byzantine mosaic decoration demus

OTTO DEMUS

BYZANTINE MOSAIC

DECORATION

ASPECTS OF MONUMENTAL ART

IN BYZANTIUM

BOSTON BOOK & ART SHOP

Boston, Massachusetts

I

THE CLASSICAL SYSTEM OF MIDDLE BYZANTINE

CHURCH DECORATION

INTRODUCTION

ONLY within comparatively recent times have historians of Byzantine

art, abandoning a purely archaeological and iconographical approach

to their subject, begun to consider the monuments it has left primarily

on their merits as works of art. The formal qualities of each image, the stylistic

texture of each figure, have at last become their main centre of interest. These researches, in which Russian and American scholars are especially active, arc

encroaching more and more upon a branch of study which, until lately, had

prided itself above all else on the closeness of its relationship with exact arche

ology, philology and theology. Points of view which, not so many years ago,

were expressed only by word of mouth or in the informal lecture, are now

finding expression in print, and arc gradually transforming our whole attitude

to Byzantinism.

The younger generation of art historians in this field have devoted most of

their labours to analysing what might be called the ** microcosm" of Byzantine

art; the analysis of its macrocosmic style, on the other hand, has been almost

entirely neglected. Yet in the case of an art which has left us some of the

grandest and most homogeneous of decorations, this aspect is deserving of

special study. If they arc considered as isolated works, Byzantine monumental

paintings lose something of their essential value. They were not created as

independent pictures. Their relation to each other, to their architectural frame

work and to the beholder must have been a principal concern of their creators.

In the case of church decoration—the field in which Byzantine art rose, perhaps,

to its greatest heights—the single works arc parts of an organic, hardly divisible

whole which is built up according to certain fixed principles. In the classical

period of middle Byzantine art—that is, from the end of the ninth to the end

of the eleventh century—these principles seem to form a fairly consistent whole,

in which certain features are permissible and even necessary, while others, con

sidered out of keeping with them, are avoided. This system was not purely a

formalistic one; it was the theologian's concern as much as the artist's. But its

iconographical and its formal sides are but different aspects of a single under

lying principle which might be defined, crudely perhaps, as the establishment of

BYZANTINE MOSAIC DECORATION

I an intimate relationship between the world of the beholder and the world of the

image. This relationship was certainly closer in Byzantine than it was in

Western mediaeval art. In Byzantium the beholder was not kept at a distance

from the image; he entered within its aura of sanctity, and the image, in turn,

partook of the space in which he moved. He was not so much a "beholder"

as a "participant*5. While it does not aim at illusion, Byzantine religious art

abolishes all clear distinction between the world of reality and the world of

appearance.

The complete realization of the formal and iconographic scheme which grew

out of this fundamental principle is, however, an ideal or, at least, an optimal

case. The nearest approach to this ideal, the classical solution, is embodied in

the mosaic decorations of the great monastic churches of the eleventh century.

The principles followed in these monuments of Imperial piety and munificence

differ widely from those which underlie early Christian and pre-Iconoclast

Byzantine, and still more Western mediaeval decorations.

y The first thing which strikes the student of middle Byzantine decorative schemes

is the comparatively narrow range of their subject-matter (i, 42, 43), They show

a lack of invention and imagination all the more remarkable when we realize

that there existed at the same time in Byzantium a powerful current of highly

imaginative art which had its source in the naive imagery of the people. But

this current seems to have found expression not so much in monumental painting

(save in the provincial hinterland) as in the illustration of popular religious

literature, homiletdc or allegorical, even of Scriptural books such as the Psalter

or liturgical compositions such as the Akaihistos. In illustrating such texts as

these the miniaturists could draw on the store of antique, sub-antique and

Oriental imagery which lent itself to an associative elaboration of the written

. word. No such freedom was either claimed by or permitted to the artists who,

as the representatives of official hieratic art, adorned the mosaic-decorated

churches of the Byzantine middle ages. The moralistic vein which so greatly

influenced the decoration of Western cathedrals, with their didactic and ethical

cycles, was likewise entirely outside the Byzantine range. The occupations and

labours of the months, for instance, the personified virtues and vices, the alle

gories of the liberal arts, the expression of eschatological fears and hopes, all that

makes up the monumental speculum universale of Western decorations,1 we shall

look for in vain inside the magic circle of middle Byzantine mosaic compositions.

These latter are to be taken as the Byzantine Church's representation of itself

; rather than of Greek or Eastern Christianity; as the product of abstract

theolog\r rather than of popular piety. There is nothing original, nothing

individual, about middle Byzantine decorations if they are considered from the

Western point of view, that is, with regard to their contents. The individual

pictures do not aim at evoking the emotions of pity, fear or hope; any such appeal

would have been felt as all too human, too theatrical, and out of tune with the

THE CLASSICAL SYSTEM

tenor of religious assurance which pervades the ensembles and leaves no room

for spiritual and moral problems. The pictures make their appeal to the

beholder not as an individual human being, a soul to be saved, as it were, but

as a member of the Church, with his own assigned place in the hierarchical

organization. The stress is not laid on the single picture in isolation: that is

"common form" to the beholder, since it follows a strict iconographic type, like

the suras of the Koran in Islamic decoration, which all the faithful know by

heart. The point of interest is rather the combination of the single items of the

decoration, their relationship to each other and to the whole. It is in this

arrangement that we must look for the unique achievement of middle Byzan

tine decoration. The single pictures were more or less standardized by

tradition; the ever-new problem for the theologian and for the artist was the

building up of the scheme as a whole. This is true not only of the content of

the pictures, but also of their visual qualities. In the latter respect it involves

a mannerist approach to forms in so far as figures, pictures and ensembles are

built up out of traditionally fixed details and units; for the content, it involves

a preponderance of the systematic, the sociological interest in relations rather

than a preoccupation with problems of ethics.

In these schemes of decoration all the parts are visible to the beholder,

unlike those Western mediaeval, especially Gothic, decorations, of which some

essential constituents, once they were finished and set up in their inaccessible

places, could never be seen by human eye. A majestic singleness of purpose

runs right through the Byzantine schemes. Their authors seem to have had as

their main aim to represent the central formula of Byzantine theology, the

Christological dogma, together with its implications in the organization and the

ritual of the Byzantine Church. There are no pictures which have not some

relation to this central dogma: representations of Christ in His various aspects,

of the Virgin, of Angels, Prophets, Apostles and Saints arranged in a hierarchical

order which also includes temporal rulers as Christ's vicegerents on earth.

Historical cycles and subjects from the Old and the New Testaments, or from

apocryphal and legendary writings, are inserted in this hierarchical system not

so much for their independent narrative value as for their importance as testi-.

monies to the truth of the central dogma.

THE THEORY OF THE ICON

Every single picture, indeed, is conceived in this sense, and middle Byzantine

pictorial art as a whole draws its raison d'etre from a doctrine which developed

in connection with Christological dogma. This doctrine was evolved during the

Iconoclastic controversy of the eighth and ninth centuries.2 The relation

between the prototype and its image, argued Theodore of Studium and John of


Damascus, is analogous to that between God the Father and Christ His Son.

The Prototype, in accordance with Neoplatonic ideas, is thought of as producing

its image of necessity, as a shadow is cast by a material object, in the same way

as the Father produces the Son and the whole hierarchy of the invisible and the

visible world. Thus the world itself becomes an uninterrupted series of "images"

which includes in descending order from Christ, the image of God, the Proorismoi

(the Neoplatonic "ideas'.'), man, symbolic objects and, finally, the images of

the painter, all emanating of necessity from their various prototypes and through

them from the Archetype, God. This process of emanation imparts to the image

something of the sanctity of the archetype: the image, although differing from

its prototype kot* oucriav (according to its essence), is nevertheless identical with it koc6' CnrroorcKnv (according to its meaning), and the worship accorded to the

image (xrpoaKOvriais Tim-paK^) is passed on through the image to its prototype.

The Chiistological theme, however, dominated the doctrinal basis of Byzan

tine theory regarding images not only per analogiam but also in a more direct

manner. One of the arguments against pictures and statues put forward by the

Iconoclasts had been that any representation of Christ was impossible, since

every representation (-rrepiypa^) must either depict Him as a mere Man, thereby

denying His Godhead and falling into the anathematized error of Nestorius; or

with His two natures, divine and human, intermingled (x^c-is), thus following

the heresy of Eutyches. The charge of heresy, however, was returned by the

Iconodulcs, who maintained not only that it was possible to represent Christ

without falling into heresy, but that denial of this possibility was itself a heresy.

Christ would not have manifested Himself in human form if that form were

indeed unfit to receive and express the Divine nature. To deny that He could

be represented in the form He took in His Incarnation was to doubt the Incarna

tion itself and with it the redeeming power of the Passion. The Incarnation

could not be considered complete, or Christ's human nature genuine, if He were

not capable of being depicted In the form of man. The fact that a picture of

Christ can be painted furnishes a proof of the reality and completeness of His

Incarnation.3 A painted representation of Christ is as truly a symbolic repro

duction of the Incarnation as the Holy Liturgy is a reproduction of the Passion.

The latter presupposes the former, and the artist who conceives and creates an

image conforming to certain rules is exercising a function similar to that of the

priest.

Three main ideas of paramount importance for the whole subsequent history

:>f Byzantine art emerge from this reasoning on the doctrine of images. First,

the picture, if created in the "right manner", is a magical counterpart of the

prototype, and has a magical identity with it; second, the representation of a

noly person is worthy of veneration; thirdly, every image has its place in a

:ontinuous hierarchy.

To achieve its magical identity with the prototype, the image must possess


"similarity" (toutottis ttjs opotcbaEcos). It must depict the characteristic features

of a holy person or a sacred event in accordance with authentic sources. The

sources were either images of supernatural origin (ccxeipo-rroiriTa), contemporary

portraits or descriptions, or, in the case of scenic representations, the Holy

Scriptures. The outcome was a kind of abstract vcrism, governed by a sacred

iconography which laid down, enforced and preserved certain rules. In the case

of representations of holy persons, this vcrism made for portraiture in thesense

of attaching distinguishing features to a general scheme of the human face and

form; in that of scenic representations, for plausibility in the rendering of an

action or a situation. If this was done according to the rules the "magical

identity" was established, and the beholder found himself face to face with the

holy persons or the sacred events themselves through the medium of the image.

He was confronted with the prototypes, he conversed with the holy persons, and

himself entered the holy places, Bethlehem, Jerusalem or Golgotha.

The second idea, that of the vcncrability of the icons, follows logically from

that of magical identity.'£ 'The image is not a world by itself; it is related to the beholder, and its magical identity with the prototype exists only for and through

him. It is this that distinguishes the icon from the idol. To establish the rela

tion with the beholder, to be fit to receive his veneration, the picture must be

visible, comprehensible, easy to recognize and to interpret. Single figures must

be identified cither by unmistakable attributes or by an inscription. So that

they may receive their due veneration from the beholder they must face him,

that is, they must be represented in frontal attitude; only so do they converse

fully with the beholder (2). In a scenic image, which likewise must be charac

terized by an inscription (to fix its CnTocrraats or meaning, which in this case is

not a person but an event), everything must be clear for the beholder to perceive.

Details must not detract from the main theme; the principal figure must occupy

the most conspicuous place; meaning, direction and result of the action must be

plainly shown; actors and counter-actors must be separated into clear-cut groups.

The compositional scheme which best answers these demands is the symmetrical

arrangement, which at the same time is in itself the "sacred form" par excellence.

Frontality, however, cannot always be achieved in scenic representations:

its rigid observance by all the participants in a scene would make the rendering

of an event or an action all but impossible. No active relationship between the

figures could be established under such a limitation, and the law of plausibility,

the demand for authenticity, would thus be violated. This was indeed a

dilemma for an art which did not know or at any rate recognize pictorial space.

Apart from spatial illusionism, the most natural way of rendering an active

relation between two or more figures on a fiat surface would have been to repre

sent them in strict profile. The figures would then have faced each other, their

looks and gestures would have seemed to reach their aims. But this would have

severed their relation with the beholder.5 The attempt was indeed made in


such scenes as the Annunciation, the Baptism, the Transfiguration, the Entry

into Jerusalem, the Crucifixion, the Doubting of Thomas and the Ascension—

scenes in which action counts for less than the representation of glorified existence

—to depict at least the main figures in frontal attitudes. But in other scenes,

where action is the main theme, this was impossible. For such cases, and for

almost all the secondary figures in scenic representations, Byzantine art made

use of a compromise between the attitude appropriate to action, the profile, and

the attitude appropriate to sacred representation, the full face. The three-

quarter view, combining both attitudes, was introduced; and this even became

the dominant mode of projection in Byzantine art. Its ambivalent character

allows of either interpretation; within the picture as a profile, in relation to the

beholder as a frontal view.

In this system there is hardly any place for the strict profile; a figure so

represented has no contact with the beholder. It is regarded as averted, and

:hus does not share in the veneration accorded to the image. Consequently,

n the hierarchical art of icon painting, this aspect is used only for figures which

•epresent evil forces, such as Satan at the Temptation, Judas at the Last Supper (3)

ind the Betrayal. From the point of view of form, the face drawn in strict

profile is for the Byzantine artist only half a face showing, as it does, only a single

rye. It is drawn exactly like a face in three-quarter view in which the half-

iverted side has been suppressed. This method of constructing a profile gives

he face a curious quality of incompleteness (23). Formally, something is missing—

ust as the otherwise indispensable relation to the beholder is left out as regards

he meaning. But the evil figures must not receive the venerating gaze of the

>eholder, and they themselves must not seem to be looking at him: iconographic

heory and popular fear of the "evil eye" go hand in hand. Outside the strictest

chool of Byzantine iconography the pure profile is also, though seldom, used for

econdary figures. Full back views do not occur at all in the classical period of

niddle Byzantine art; for to the Byzantine beholder such figures would not be "present" at all.

As a result, the whole scale of turning is toned down in classical Byzantine

rt. It is as if the figures were somehow chained to the beholder; as if they

/ere forced as much as is compatible with their actions into frontal positions.

The generally lowered key gives, on the other hand, a heightened importance d the slightest deviations from strict frontality. The eye, expecting frontal

ttitudes, registers deviations in posture and glance much more strongly than it 'ould if frontality were the exception, as it is in Western art. The projection

sed in scenic images is, from the formal aspect, a qualified en face rather than real three-quarter view.

But even this three-quarter view, apparently, did not seem to the Byzantine

rtist an entirely satisfactory solution. The gestures and gaze of the figures

ill miss their aims: they do not meet within the picture, half-way between

8


figures engaged in intercourse, but in an imaginary point of focus outside, that is, in front of it. There is a dead angle between the actors in a scene, an angle which is not quite bridged even by oblique glances. The action takes' on a stiff frozen air. To remedy this, to give plausibility and fluency to the representa tion, two correctives were applied, at first separately, in two different realms of Byzantine art, but from the twelfth century onwards more or less indiscrimi nately. On flat surfaces, especially in miniatures, ivories, and the like, move ments and gestures were intensified in order to bridge the gap between the figures as the actors in the scene. In a field of art which made use of neither pictorial space nor psychological differentiation, gestures and movements could be intensified only, so to speak, from outside, by a heightening of tempo. Intensity of action was preferably conveyed by locomotion. The figures run towards each other with outstretched hands and flying garments (3). The Angel

of the Annunciation rushes towards Mary, the Baptist climbs the bank of Jordan with hasty strides, executioners fell their victims in full career, martyrs re ceive the death-stroke as they fell (61b). These are not the baroque exaggerations of a late phase, but attempts to establish the necessary contacts and to save the picture from appearing as a disconnected row of single figures. There is a definite tendency in this method of rendering action to point forward in time, to make the result of the action apparent together with the action itself, and so not only to connect the figures of one picture among themselves, but also to establish a relation between the successive pictures of a narrative cycle.

This remedy, however, satisfactory and fertile as it was in illustrative pictures of small size, was hardly applicable to monumental paintings on the grand scale. The violent movements would have seemed too undignified, the whirling forms too contorted and complicated. Another means was therefore needed by the Byzantine decorators to bridge the dead angle and save the threatened coherence. The solution they found was as simple as it was ingenious. They placed their pictures in niches, on curved surfaces. These curved or angular surfaces achieve what an even, flat surface could not: the figures which on a flat ground were only half-turned towards each other could now face each other fully without having ̂ to give up their dignified frontality or semi-frontality. Painted on opposite sides of curved or anguiar niches, they are actually facing each other m real space, and converse with each other across that physical space which is now, as it were, included in the picture. The curvature in the real space supplies what was lacking in the coherence of the image (4).

The firm position of the painted figures in physical space makes spatial symbols in the picture itself unnecessary. No illusion is needed in pictures which enclose real space, and no setting is required to clarify the position of the figures. The whole of the spatial receptacles (such the pictures really are) can be devoted to the figures themselves and to such motives as are required from the icono graphic point of view. Restrained gestures and movements are sufficient to


establish the

necessary contact.

A large

part of the g

olden ground can be left

empty, surrounding the figures w

ith an aura of sanctity.

This golden ground in

middle Byzantine

mosaics is

no: a s

ymbol of unlimited space;

it need not

be pushed back, as it were, in order to leave sufficient space for the figures to act.

They move and gesticulate across the physical space w

hich opens u

p in front of

the golden

walls. The shape

and the

confines of this

physical space

arc not

dissolved, but rather stressed and clarified,

by the solid coating of gold.

The

setting of the gold

is close and firm,

producing a metallic surface

whose high

lights and shades bring out the plastic shape of the niche.

There is n

o need, in this formal system, for the figures engaged in intercourse

of whatever kind to approach close to each other. On the contrary, they h

ad to

be placed at s

ome distance apart in order that they might be brought opposite

each other b

y the curving of the ground.

The resulting distances a

nd empty

spaces are filled

with a

tension, an air

of expectancy,

which makes the

event depicted

even more dramatic

in the

classical sense

than violent

action and

gesticulation, or a closely knit grouping, could h

ave made it.

The casura con

tribute also to the legibility, to the plausibility of the image.

The main figure

is clearly discernible, because comparatively isolated, and presents itself unmis

takably as t

he main object o

f veneration.

But the venerability of the icon did not affect its composition alone; it also influenced the choice of material.

Controversy about the "matter" (OAti) of the images played a large part in the Iconoclastic struggle.

It was but natural that,

to counter the a

rguments of the Iconoclasts r

egarding the incongruity o

f repre

senting the Divine in common and cheap material, the Iconodules should h

ave

chosen the m

ost precious

material for

this purpose.

Mosaic, with

its gemlike character a

nd its profusion of gold, m

ust have appeared, together with enamel,

as the

substance most worthy of b

ecoming the

vehicle of divine

ideas. It

is partly for this

reason that mosaic played so

important a part in the evolution

Dt post-Iconoclastic painting, and indeed

actually dominate-1 it.

It allowed of

Dure and radiant colours w

hose substance h

ad gone through the purifying e

lement

)f fire and which seemed most apt to represent

vhe unearthly splendour of the divine prototypes.

ARCHITECTURAL AND TECHNICAL CONDITIONS

These prototypes themselves,

to the Byzantine mind, stand to e

ach other in

i hierarchic relation, and so their i

mages must express this relationship.

They

nust occupy their d

ue place in a

hierarchy of values in w

hich the i

mage of the

vll-Ruler occupies the central

and most elevated

position. Clearly,

a hierar-

hical system of images based on the principles w

hich governed the Byzantine

Church's own organization could

be fully expressed only through an architec-

10


tural framework that

furnished a hierarchy of

receptacles with

in which the

pictures could be arranged. A purely narrative

sequence of pictures,

in the

Western sense,

or a didactic s

cheme could be displayed o

n almost any surface in

almost any arrangement. Whether it

was used

to decorate portals,

facades,

interior walls or stained-glass w

indows did not greatly matter.

But a Byzantine

programme always needed a special

framework, namely that

in which it

had

grown up, a

nd which it w

as developed to suit.

This framework was the classical

type of middle Byzantine ecclesiastical

architecture, the cross-in-squarc

church

with a central cupola.6

The shaping of this architectural

type was a lengthy process,

and the final

solution was arrived at b

y several c

oncurring paths.

The essential

idea seems

to have been conceived as

early as

the sixth century.

Architects with widely

different traditional backgrounds approached the p

roblem from different sides:

from the

centralized plan of the

free cross,

the octagon,

the basilica

with a

central cupola, the

semi-basilican plans of Salonica

and Ancyra, etc.

Tem

porary solutions w

hich lay in the g

eneral direction o

f development, such as the

Constantinopolitan type

with five

naves, the

ambulatory church,

or the

tri-conchos, were tried out, as it w

ere, and finally a

bandoned for the classical

type

of the cross-in-square.

The history

of this

development allows

of an almost

teleological interpretation.

There is e

vidence of a

conscious-search for a final

solution in accord with the liturgical n

eeds and the aesthetic ideals of

the time.

Local differentiations g

ave way before the quest for this ideal

type; an

d, when

finally elaborated, it was never abandoned, and remained the basis of the w

hole

of the subsequent development. Even changes of scale

did not greatly

affect the d

ominant idea.

The final type, fully evolved b

y the e

nd of

the ninth century, was something strangely perfect, s

omething which, from the liturgical a

nd from

the formal points o

f view, c

ould hardly be improved upon.7

This high perfec

tion might have resulted in sterility, h

ad not the central architectural idea b

een

flexible enough to

leave room for variation.

The plan was,

in short,

that of a

cruciform

space formed by the

vaulted superstructure

of transepts a

rranged crosswise a

nd crowned in

the centre by a

higher cupola.

The angles

between the

arms of the

cross are

filled in

with lower vaulted

units, producing a full s

quare in the g

round-plan but preserving

the cross-shaped

space in

the superstructure.

Three apses

arc joined

to

the square on the

east and >si

entrance hall

(sometimes two) stands

before it on

the west. ^

The most characteristic

feature of this

architectural scheme

is its

elasticity, inherent in the fact that the spatial conception finds its complete fulfil

ment in

the vaulted superstructure o

f the building: it is a

conception expressed

in vaults,

and not in the e

lements of the ground-plan.

The latter,

indeed, can

vary a good deal without affecting t

he guiding idea.

The cross w

ith the central

cupola can be imposed on almost any ground-plan, even on that o

f a basilica;

or rather almost any ground-plan can be subjoined to

the ideal

system of

n


vaulted space. The configuration of the vaults itself remains invariable; it is the image of the changeless and perpetual celestial world set over against the varying earthly sphere of the ground-plan which was conditioned by terrain, space, special purpose and changing fashion. However different the ground-plan (the space in which the beholder moves) may have been, his upward look always meets the familiar configuration of golden vaults (in what may be called the optical space). The cupola always dominates the impression. Even the modern beholder directs to it his first glance. From the cupola his eye gradually descends to the horizontal views.

This process of successive apperception from the cupola downwards is in complete accord with the assthetic character of Byzantine architecture: a Byzan-dne building does not embody the structural energies of growth, as Gothic irehitecture does, or those of massive weight, as so often in Romanesque build-ngs, or yet the idea of perfect equilibrium of forces, like the Greek temple Byzantine architecture is essentially a "hanging" architecture; its vaults depend rom above without any weight of-their own. The columns are conceived esthetically, not as supporting elements, but as descending tentacles or hanging •oots. They lack all that would make them appear to support an appropriate veight: they have no entasis, no crenellations, no fluting, no socles; neither does he shape of the capitals suggest the function of support. This impression is not xmfmed to the modern beholder: it is quite clearly formulated in contemporary iyzantine ekfihraseisJ The architectonic conception of a building developing lownwards is in complete accord with the hierarchical way of thought mani* ested in every sphere of Byzantine life, from the political to the religious, as it is o be met with m the hierarchic conception of the series of images descending rom the supreme archetype.

^ The cross-in-squarc system of vaults is indeed the ideal receptacle for a jerarchical system of icons. Each single icon receives its fitting place according ■> its degree of sanctity or importance. Just as Byzantine architecture is pri-larily an architecture of vaults, and is therefore concerned especially with the pper parts of the structure, so Byzantine decoration is primarily at home in the aulted parts of the church. The lower the icons descend, the more they leave icir proper sphere. Especially does mosaic unfold its finest qualities in the aults, on curved surfaces. This is equally true of the qualities of technique >rm and hieratic content. Technically, the mosaic is more safe and solid on -Lrved surfaces, where it forms a kind of vault in itself, and, as has become pparent during modern restorations, is often sufficiently supported by the ten on of the vault-like arrangement of the cubes on the curved surface even when icy have become detached from the real vault by the corrosion of the mortar at joins them. On flat surfaces, on the other hand, there is no such tension ) mutual support between the cubes; loose parts crumble and fall off. This' in be seen in almost all mosaics; the curved parts are generally better preserved

12


than the fiat ones. Mosaicists who had to work on flat surfaces in fiat for instance, even went so far as to curve the setting-bed slight or at ^ incline it out of the vertical and to undulate the plane surface » in < T* t0 provide some sort of tension and to bring out the formal qualities F I to too are greatly enhanced by curved or undulated surfaces. Only on sZh ^ faces can the sparkling of the enamel cubes and the glitter of the sold bb out as essential factors. Just as stained glass must be seen .gffi£fe so a mosaic displays its peculiar charm under the play of £onS ght' surface On fiat walls mosaics look comparatively dulh dthTdS tl 3T? dark and lustreless, or the golden ground is lit up while the figuS^S o Uy dark and colourless silhouettes (5). 6 stand °ut as

THE ICON IN SPACE

it* It° ̂ i** mlddle B):2antine mosaic Painting came to maturity, and its proper place, on curved surfaces also influenced the shape of the cX the way in which they were set. These surfaces did not admit of the &e large, irregular pieces of stone and enamel which were employed for^ Christian wall mosaics, in which each tessera stood for one dLtfect patch colour and took the shape of that patch. The curved receptacle? of S Byzantine mosaics called for small uniformly shaped cubes, &LdZl close network whose lines run in form-designing curves. The tenskS of curves mirrors the tension of the rounded receptacles. In this ™£ the

To describe these mosaics, encased in cupolas, apsides, squinches

?5 ̂ nkhCSK ? flat' °r two-dune^ioAalf would le ̂ ap True, there is no space behind the "picture-plane" of these mosaics. But

ITS** • »f ^ ̂ end0Sed ^ ±e nich^ in front; ^d this spacc ncluded m the picture. The image is not separated from the beholder by ! imagmary glass pane" of the picture plane behind which an nj££

tare begins: it opens into the real space in front, where the beholder lives moves. His space and the space in which the holy persons exist and aa xdentical, just as the icon itself is magically identical with the holy perSn sacred event The Byzantine church itself is the "picture-space" rfS It is the ideal iconostaas; it is itself, as a whole, all icon giving «alSr conception of the dzvme world order. Only in diis mediuS wWch b

i ̂ ST"" ̂ ̂ *****" Can the la£ter fed ** he » hinS holy events and conversing with the holy persons. He is not

SunH hT! tC " b0dlly ?ndTd in thc ^d kon of ̂ <*"«*; £ rounded by the congregation of the saints and takes part in the events he

13

This feeling of the reality of space developed in Byzantine painting at the

time when Byzantine sculpture died, as a result of the Iconoclastic controversy.

Much of the spatial quality of sculpture in the round, which may be said to

exist in the same space as the beholder, went into the development of Byzantine

painting. Byzantine monumental painting was, indeed, the legitimate successor

of monumental sculpture.

If, however, the icons were to exist in, and to share, a space which is normally

the domain of the beholder, it was more than ever necessary to place them in

individual receptacles—in spatial units which are, as it were, excrescences of

the general space. Moreover, since the images are not links in a continuous

chain of narrative, they must not flow into one another: they must be clearly

separated and each must occupy its own place in the same manner as the events

and persons they represent occupy distinct places in the hierarchical system.

The formal means to this end is the separate framing of each single receptacle.

The single units are set off either by their characteristic shapes as spatial units,

especially in the upper parts of the building, or, in the lower parts, by being

embedded separately in the quiet colour foil of the marble linings. This marble

entablature with its grey, brown, reddish or green hues covers practically all the

vertical surfaces of the wails in middle Byzantine mosaic churches, leaving for

iie mosaics only niches in which they are placed like jewels in a quiet setting.

Nothing is more alien to the monumental mosaic decorations of these churches

n the central area than the almost indiscriminate covering of the walls with

"nosaic pictures which is found in the twelfth century in Sicily, Venice and other

:olonial outposts of Byzantine art. In Byzantium itself the mosaics never lose

:he quality of precious stones in an ample setting. The icons never cease to be

ndividually framed spatial units; their connection with one another is estab-

ished not by crowded contiguity on the surface but by an intricate system of

•elations in space.

THE IDEAL 1C0NOGRAPHIC SCHEME OF THE

CROSS-IN-SQUARE CHURCH

These relations were governed, in the classical period of the tenth and

leventh centuries, by formal and theological principles. There have been

.nalysed and interpreted by a number of scholars, above all by Gabriel Millet.10

7he methods used are sometimes open to criticism, as when, in order to recon-

truct the "ideal51 scheme of middle Byzantine church decoration, observations

rawn from monuments of widely different dates are combined with others

leaned from such varying sources as ihe writings of iconologists and icono-

raphers from Theodore Studites and John of Damascus down to Simeon of

'hessalonike and John of Euboea, the ekpkrasds of exegctes like Constantine

Rhodius, Nicholas Mcsarites, or late painters' guides such asi the meneia. For scholars to seek for an actual monument embody ng scheme^ recalls Goethe's quest of the UrPfianZe. But it is P^^fJSttaJ principles if that phrase be taken to covCr not only theological and oxle«s conceptions differing according to date, locality and speaal desunauon, b* a the adgencies of varying architectural schemes. For such an ̂ aly^tJ the particular interpretations given by different .conographen, to the parts of the Byzantine church building and to its d6^ so much as the fact that such interpretations were proffered at ̂ > ̂ ccn of them became the guiding principles of middle Byzantine hermeneuues. => from this angle, we can distinguish three systems of mterprctauonwh^ found interlinked in every Byzantine scheme of decoration of the lead, g,

"ThSyzaEe church is, fim, an image of the Kosmos, symbolizing heaven, paradise (or the Holy Land) and the terrestrial world m an ordered hierarchy, descending from the sphere of the cupolas, which represent heaven^ earthly zone of the lower parts. The higher a picture is placed in thearcW tural framework, the more sacred it is held to be. The ̂ .^^(and more specifically topographical. The building a conceived as *^nago. I so as magically identical with) the places sanctified by Christ s earthly^ affords the possibility of very detailed topographical ^^5°'^°^ » which every part of the church is identified with some place in the The faithful who gaze at the cycle of images can make a sy the Holy Land by simply contemplating the images intheir local chur£ nJ perhaps" is the reason why actual pilgrimages to Palestine played so ummpor a part in Byzantine religious life, and why there was so little response*the * of the Crusades anywhere in the Byzantine empire It may also ,acco£[ ;dnian fact that we do not find in Byzantium reproductions of "»dmdualPal^ shrines, those reproductions of the Holy Sepulchre, for instance, which important a part in Western architecture and devotional life

In the realm of topographical symbolism, however, more than ui any ot ^ over-interpretation set in fairly soon, more than one symbohc identification be S applied to one locality or even to a single object of church ̂ f^^^or in of 5ns can be found in the Ecclesiastical History of the Patriarch Germanos o the writings of Simeon of Thessalonikc." They make it clear that we arc c^ fronted in these speculations with ex post facto interpretations and^ not guiding principles. But such principles did exist, even if every single <*#* for a new adaptation of them, and they are shown in the fact that certain' P*^ the building were more or less consistently chosen for the location 01

icons. . , , ty.~ Calendar of Th* third kind of symbolical interpretation was^based on "JcOaK

the Christian year." From this point of view, the church is an image

15


estival cycle as laid down in the liturgy, and the icons are arranged in accordance vith the liturgical sequence of the ecclesiastical festivals. Even the portraits of he saints follow to some extent their grouping in the Calendar, and the arrange-nent of larger narrative cycles is frequently guided by the order of the Peri-;opes, especially as regards the scenes connected with Easter. Thus the images

ire arranged ia a magic cycle. The relationship between the individual scenes las regard not to the "historical" time of the simple narrative but to the "sym-

>olic" time of the liturgical cycle. This cycle is a closed one, repeating itself :very year, during which, at the time of the corresponding festival, each image n turn comes to the front for the purpose of veneration, to step back again into ts place for the rest of the year when its magic moment has passed. The pro-:>und contrast between this conception of time and that implicit in Western

lecorative schemes is obvious: in the latter a series of scenes illustrates an histori-al sequence of events, with its beginning and end clearly marked and with a lefinite direction parallel with the unrolling of the story. In the strict arrange-icnt of Byzantine decorations the time element is symbolical; it is interlinked nth the topographical symbolism of the building, and therefore closely con-ected with the spatial clement. The flow of time is converted into an ever-xurring circle moving round a static centre. These two conceptions of time Drrespond to the two dominant architectural types: the Western to the basilican /pe," with its rhythmic movement from entrance to apse, from beginning to id; the Byzantine to the domed centralized building which has no strongly nphasized direction, and in which the movement has no aim, being simply a rcular motion round the centre.

All three Byzantine systems of interpretation, the hierarchical cosmic, the >pographical and the liturgico-chronological, are so closely accommodated to ic dominant architectural type of the cross-in-square church that they must, fact, have been elaborated for such a building. Only within this framework

>uld a scheme devised after these principles be satisfactorily placed. Every tempt, therefore, to adapt such a programme to other types of architecture ust have met with great difficulties, and must consequently have resulted in a wakening of the original concepts, as can actually be seen in the provinces.

THE THREE

The most obvious articulation to be observed in a middle Byzantine mosaic coration is that which corresponds to the tripartition into heaven, paradise or Dly Land, and terrestrial world. Three zones" can be clearly distinguished' st, the cupolas and high vaults, including the conch of the apse; second, the Lunches, pendentives and upper parts of the vaults; and thirdly, the lower or :ondary vaults and the lower parts of the walls. These three zones *re, in

16


most cases, separated by plastic cosmetes—narrow bands of carved stone or stucco which run round the whole edifice.

§ Cupolas and Apses

The uppermost zone, the celestial sphere of the microcosm of the church,

contains only representations of the holiest persons (Christ, the Virgin, Angels)' and of scenes which are imagined as taking place in heaven or in which heaven is either the source or the aim of the action depicted. Byzantine art from the

ninth to the end of the eleventh century made use of only three schemes of cupola de

coration: the Ascension (6), the Descent of the Holy Ghost (8), and the Glory of

the Pantocrator (7), the All-Ruler. This peculiarity distinguishes the strict scheme

of the Middle Ages from early Byzantine as well as from Italo-Byzantine decora tion. In the five cupolas of the Justinianic church of the Apostles in Constanti nople,17 for instance, there had been five different representations, each forming

part of the narrative cycle which filled the whole church. After the Icono

clastic controversy, however, and in connection with the subsequent emergence

of the symbolic interpretation of the church building, the cupolas were strictly sti apart from the narrative cycle. From the ninth century onwards they con

tained only representations in which the narrative character had been displaced entirely by the dogmatic content. The three themes above-mentioned domi nated Byzantine cupola decorations after the Iconoclastic controversy to such

an extent that others were scarcely thinkable; even the small cupolas of entrance

halls were decorated with them. Examples can be found, for instance, in Chios,18 and later in the Kahrieh Djami in Constantinople,1* where the scheme

of the Pantocrator cupola was adapted to the special contexts,20 and where Mary has supplanted the central figure of Christ.

The Ascension, Pentecost and Pantocrator themes fulfil, moreover, the formal

conditions of middle Byzantine cupola decoration in that they contain an

organic centre round which the composition could be arranged radially.

Apostles, Angels or Prophets arc placed like the spokes of a wheel round the

medallion with the ascending Christ, the Hetoimasia (prepared throne), or the

bust of the Pantocrator respectively in the centre. But the decoration of a

middle Byzantine cupola was not conceived, primarily, as a flat design. The dome was thought of as enclosing real space, of a bell-shaped form, and the

decoration is in reality a spatial arrangement. The standing or seated figures are quite actually standing or sitting around and below the elevated central

motif. The cavity of the cupola is a perfect equivalent of the imagined spatial

arrangement which the figures would occupy if the scenes were being enacted

in reality; it is a receptacle which exactly fits the spatial conception. Under

these circumstances there is obviously no room for illusionistic picture-space in this art: the exigencies of spatial realism are more directly fulfilled by this means

17


than they could have been by the cleverest tricks of illusionism. The figures

stand or sit at the limits of the physical space, or rather they form and underline

these limits. If there is no appreciable space behind them, behind the picture

plane, there is, on the other hand, space, the real space, in front.

A special rhythm, different in kind and degree from the rhythmic movement

of figures in a flat design, binds the figures together and connects them with the

elevated central motive. Complicated attitudes, especially in the representation

of the Ascension, create a rhythmical movement, almost a dance, round the

cupolas; and glances, gestures and intermediary motives (e.g. the trees in the

Ascension and the rays of light in the Pentecost) lead upward to the centre.

This formal conception, however, is not radially symmetrical in the strict sense.

The close relationship between icon and beholder enforces a definite orientation

in accordance with the main optical axis. The central medallion (Christ,

Hetoimasia, Pantocrator) is so arranged as to appear upright in the view from

west to east; and the lower parts of the cupola are also composed so as to fit in

with this line of vision. Not only are the main figures, iconographically speak

ing, assembled in the eastern hal£ so that they look westwards and confront the

beholder in an almost frontal attitude; they are, in addition, arranged sym

metrically along the middle axis of this view, offering to the beholder a hieratic

icon: Mary with Angels and with Peter and Paul in the Ascension, Peter and

Paul frontally enthroned in the Pentecost, and frontal Archangels and Prophets

in the Pantocrator cupolas. The eastern part of the cupola is, so to speak, the

facade of the pictorial edifice; it contains the quintessence of the image. The

figures in the western, the averted, half of the cupola, on the other hand, are

shown in more agitated and less frontal postures, even positions which, in the

lowered key of Byzantine projection, appear almost back views. In the Ascen

sion these figures are looking upwards at a steep angle, making this part of the

cupola appear steeper than the seemingly flatter facade which faces the beholder.

The spatial structure of the composition is thus built up to suit the beholder who

faces east and looks up into the cupola. To such a beholder the cupola would

appear steeper in the western half and flatter in the eastern. If he stands under

the western foot of the cupola, he has almost to turn round and look upwards at

a. steep angle in order to get a view of the "averted" side; his physical and

spiritual attitude is mirrored in the postures of the figures placed in this part of

the cupola or the drum. The structure of the whole cupola, which in reality is

regular and radially symmetrical, thus appears to be asymmetrical in accordance

with the dynamic of the beholder's view.

One difficulty, however, always faced the Byzantine artist with regard to the

central motif in the zenith of the cupob.. A full-length figure placed there could

not be represented as upright without the help of illusionistic perspective,

unknown to Byzantine painting. An unforeshortened figure would necessarily

have appeared to the strict spatial feeling of the Byzantines as lying face down-

18


wards. This resulted in a certain awkwardness, and the attempt was made to

resolve it by connecting the central figure of the Ascension, for instance, with the " upright>J eastern half of the cupola by inserting a pair of flying angels sup

porting the enthroned Christ from below, and surrounding the central figure with a circular halo as if with a medallion frame. This frame softened the

realistic implications of the figure, as it were, making it appear as an axially

symmetrical motif. But even so this solution was apparently felt to be unsatis

factory, and this was perhaps one of the reasons why, in the classical period of

middle Byzantine decoration, the Ascension as a cupola composition was soon

abandoned in favour of the Pantocrator scheme, which, with a half-figure

enclosed in a medallion, offered no such difficulties. The abstract symbol of the Hetoimasia in the Pentecost could be arranged satisfactorily from the beginning.

The iconic art of cupola decoration was the most magnificent achievement

of middle Byzantine monumental painting. It was the fullest realization of the

idea of the monumental icon, conceived as a configuration in real space. The

figures move and arc arranged in this spatial receptacle according to the

dynamics of the beholder's vision. This solution—that of making the physical

space the action-space of the figures—was seldom employed either before the

ninth or after the thirteenth century. The conception of space and picture

coinciding with it is the basic principle of classical middle Byzantine art. The art

of the West made use of this conception of real space only for a short time,

about 1500, when cupola decoration became once more a main problem of

monumental painting. Leonardo in his Sala delle Asse and Correggio in his

decoration of the Abbess's chamber at Parma tried to stress the confines of the

actual space by encasing it with forms whose position in imaginary space co

incided with their situation in physical space. But very soon the Western idea

of spatial illusion led to an imaginary widening of the physical space by windows

opening into unlimited pictorial space- In the Byzantine decorations of the

classical period this never happened. The enclosed space, the confines of the

spatial icon, were never pierced by illusionistic means. The icon always kept

its well-defined shape.

The three themes of classical Byzantine cupola decoration—Ascension,

Pentecost and Pantocrator—were hardly ever represented in three cupolas of the

same building, except in monuments of "coionial" character, such as St. Mark's,

Venice21 (6, 8,9), or except as a result of later changes, as in the Holy Apostles

at Constantinople.22 The reason for this is that, in the main line of Byzantine

evolution, the cupola scheme of the Pantocrator" replaced that of the Ascension.

The theological content of the new scheme must have been originally very

similar to that of the older. It depicted Christ in Glory after the Ascension.

The feeling that the two themes are identical was so strong that they were not

represented together in a single scheme. In Pantocrator churches, for instance

in Hosios Lukas, the Ascension was even omitted in the scrips of pictures recalling


the Church festivals.

In the outposts a

nd provincial b

ackwaters of the e

mpire,

however, the

Ascension cupola scheme lived

on for

some time,

especially in

Cappadocia," Italy" and Russia.26

The Ascension cupola was connected, on

the whole, with a more narrative c

haracter in the w

hole programme.

It is thus

essentially pre-Iconoclastic,

in contrast

to the abstract

and dogmatizing repre

sentation of the Glory of the

PantocraiOr which became the natural centre of

hieratic post-Iconoclastic

decorative programmes." The two cupola schemes

excluded each other.

To use t

hem side b

y side in the s

ame church, in v

iew of

their almost identical

content, would have been a kind of pleonasm; it w

ould

also have involved confusion between two different

lines of thought. The

decoration, by being given two centres, w

ould have been deprived of its unity.

Whenever this

happened, as

in some cases

it actually

did, we may be certain

that the programme in q

uestion was laid d

own by provincial theologians w

ith

little understanding of the dogmatic clarity o

f metropolitan iconography.

The Pentecost,

on the

other hand, goes

well with either

of the

two main

cupola schemes.

It was never the centre o

f a whole programme, and could be

attached either

to the Ascension or to

the Pantocrator, provided there was a

second cupola to fill.

If this was the case, as in H

osios Lukas,28 the t

heme was

given in full,

above the sanctuary and the m

ain altar.

This position indicates

that it was not primarily intended as a

n illustration o

f the "historical" e

vent in

the course of a narrative cycle, o

r even as o

ne link in the cycle o

f feasts, b

ut as

an allegory of the divine inspiration o

f the C

hurch whose priests officiated b

elow.

This becomes even clearer

in those cases

(as in

the majority of the surviving

monuments") in w

hich there

is no cupola above the altar.

In these cases o

nly

an abridged version was used for t

he decoration of the presbytery-vault, showing

the central motif alone of the full

type, the Hetoimasia, as the seat a

nd source

of Divine inspiration.

With this the representation lost e

ven what was left o

f its

scenic character.

It became a symbol which, by virtue

of its

very generality,

was apt to t

ake on a new and wider meaning—even an eschatological i

mplica

tion. The two (or four)

adoring angels

almost invariably added underline the

generalizing character of the

symbol,30 which at

last became so

far removed

from an illustration

of the historical

Pentecost scene that

another icon illus

trating the

Descent of the Holy Ghost could be added to

the festival

cycle

without apparent duplication.

Pantocrator and Hetoimasia arc, in fact, parallel

results of a single

development—that from the

historic scene to

the dogmatic

symbol. A third e

xample of the same process

is the substitution o

f the Deesis

for the Last Judgment, which in its

fully developed form was not frequently

used within the framework of central Byzantine cross-in-square church decora

tion. It

is found, however, like

the Ascension in

its full

narrative version,

in

the outlying provinces such as Russia31 and Italy.32

The substitution o

f abridged symbols for c

omplete versions,

however, is

not

confined to

the classical

era of middle Byzantine painting.

It can be traced

20


similarly in western Gothic art (Pieta for the Descent f

rom the Cross, Christ a

nd

St John for

the Last

Supper, Christ

riding on the ass for

the Entry into J

eru

salem etc )

and in the

Italian High Renaissance (Santa C

onversazione for the

fuller'type of the

Macsta). Iconographically,

it means the

preponderance of

abstract over concrete, of idea over event; formally, the reduction of the compo

sition to its innermost core, its m

ost simplified form, w

hich allowed of a highly

abstract treatment, of perfect s

ymmetry and absolute clarity.

Nevertheless, it w

as on

ly for

a comparatively short t

ime that

these abstract symbols reigned s

upreme. As soon as their origin a

nd their special m

essage were

obscured or forgotten, the historical images were reintroduced, to figure side b

y

side with their s

ymbolic abbreviations.33

The narthex of the

middle Byzantine church has

its own heavenly zone,

marked by the i

mage of the Pantocrator over the m

ain entrance a

nd by one or

more cupola representations dedicated to the E

mmanuel or to the Virgin.

To

the heavenly zone of the c

hurch itself belongs also the c

onch of the m

ain apse.

In churches

of the dominant cross-in-square

type this invariably

contains the

figure of the Virgin, either seated in the type of the Paruxkranjos or standing in

the types of theOrante Platytera

or the2Hodegetria.

In the period

immediately after

the Iconoclastic'struggie

the standing

figure would seem to

have been

preferred; this may have been adopted because the prc-Iconoclastic type of the

Virgin enthroned was too o

pen to charges of idolatry f

rom the Iconoclasts.

In the course o

f the tenth or

the eleventh century, however, when all controversy

had died d

own, the iconographers a

nd the artists reverted to the seated image.

The standing figure w

as relegated..to the outlying,provinces34 and, perhaps,

to Paface' chapels,

where the

OranU was imbued with

a special meaning as

the Protectress of the ruler.35

In the period from the ninth to the e

nd of the eleventh

century Mary is

represented alone

in the golden conch of the apse.

Adoring

angels are placed in the high niches of the presbytery walls or in the vault above

the sanctuary where they s

eem to b

e connected as m

uch with the H

etoimasia as

with the Virgin. But later on, in the twelfth a

nd thirteenth centuries, they w

ere

moved into the c

onch itself, to attend o

n the enthroned Virgin.

There was however, another type

of Byzantine apsidal decoration

which

showed Christ in the c

onch (49).

This had been very frequent in pre-Iconoclasdc

times and it survived

after the

controversy in

the provinces.

In Byzantium

itself aad in

Greece this type w

as restricted

after the year 900 to longxuidina.

churches with no cupola, w

here the apse is the "highest" a

nd holiest place for

an icon

In cupola churches the

aps

e holds the s

econd place,

and was conse

quently the fitting site for the

image of the Virgin, second

in rank

to that of

Christ (10a).36

, ,

The side

apses were not subject

to such strict rules

as the

mam apse;

but

they were mostly connected with the "pre-history" of Christ a

nd the R

edemp

tion showing either M

ary's parents, J

oachim and Anna, or J

ohn the Baptist as

21


forerunner of the Messiah. It is in keeping with this customary programme that

the prothesis was interpreted by the iconographers as symbolizing the place of

Christ's birth, the cave of Bethlehem.

§ The Festival Cycle

The second of the three zones of the Byzantine church is dedicated.to the

life of Christ, to the pictures of the festival cycle. It harbours the monumental

calendar of the Christological festivals and is the magical counterpart of the

Holy Land. The cycle of feasts was gradually developed by selection from an

ample narrative series of New Testament scenes. It is very probable that the

decorations which immediately followed the re-establishment of icon worship

did not include any festival icons in the naos. But the austere ideal of the early

post-Iconoclastic period was relaxed in the course of the tenth and eleventh

centuries. Hosios Lukas, dating from the beginning of the eleventh century,

displays four festival icons in the naos, four more being exhibited in the narthex;

Chios, from the middle of the century, has eight (and six in the narthex); and

Daphni, dating from the end of the century, displays thirteen icons of the feasts in

the naos and a further Christological and Marianic cycle in the narthex (42,43)."

The growth of the festival cycle can also be followed in contemporary ecclesias

tical literature: there the number rises from seven to ten, twelve, sixteen and

even eighteen pictures, the full development being reached from the twelfth

century onwards.38 The classical cycle of the eleventh century comprised, at

least in theory, twelve feasts, the Dodekaeorta: Annunciation, Nativity, Presenta

tion in the Temple, Baptism, Transfiguration, Raising of Lazarus, Entry into

Jerusalem, Crucifixion, Anastasis (Descent into Hades), Ascension, Pentecost and

Koimesis (Death of the Virgin). To this series were frequently added, in pic

torial cycles, a few images which elaborated the story of Christ's Passion, namely

the Last Supper, the Washing of the Apostles' Feet, the Betrayal of Judas, the

Descent from the Cross and the Appearance to Thomas. Other developments

were attached to the story of Christ's infancy (the story of His parents, the

Adoration of the Magi, the Flight into Egypt, etc.) and to that of His teaching

(the cycle of the miracles and parables). These developments are easily recog

nizable for what they are, whereas the scenes from the Passion are treated as

equal to the twelve main festival images. The gradual enlargement of the

series of Christological scenes in the naos led, ultimately, to the dissolution of the

liturgical cycle of the festivals, the icons being merged in a fuller illustrative

cycle which retains only a few elements characteristic of the hierarchical pro

grammes of the tenth and eleventh centuries. Such traits, which survived into

, the twelfth and following centuries, axe the separation of the miracles from the

festival cycle even in its enlarged form, the emphasizing of the Crucifixion and

the Anastasis as the two key points of the Christological cycle,3* and a few

22


customary rules (fostered by hermeneutic speculation) affecting the distribution of the images. Apart from these survivals, the twelfth century shows a tendency to return to the narrative character40 prevalent in pre-Iconoclastic times.

This development was made easier and probably inaugurated by the fact that the compositional schemes of the Christological icons, unlike those of the cupola compositions, had not been developed within the sphere of monumental painting. They were taken over from small-scale pictures on a flat ground, from portable icons, miniatures, or narrative wall-paintings. These models had, originally, none of the characteristics of the "spatial" icon; they were merely flat compositions which had to be adapted, sometimes laboriously, to the new spatial receptacles and to the new liturgical roles which they had to play in the new contexts. They tended, consequently, to relapse into their original forms as soon as the hierarchical system had outlived its prime. Even in their adapted shapes within the hierarchical cycles these pictures retained something of their original narrative character. Only in deeply curved niches, as in the squinches of the great eleventh-century churches, did they assume some of the character

istics of compositions in physical space, though not, indeed, to the same degree as compositions in the cupolas. The niches are not closed in on all sides; being half-open, they do not completely surround and enclose the space within their curved walls. But their depth makes it possible to represent in them figures standing to the left and right of the centre as they might face each other in real space The Angel of the Annunciation, for instance, stands in actual fact oppo site to the almost frontal figure of the Virgin (4). The spatial distance between the two figures expresses the theme of the meeting of two different spheres. The spiritual distance between the Angel and Mary had often been naively expressed in flat icons by inserting a door or a column between them. In the spatial arrangement of the monumental icons this naive symbolism was unnecessary:

the spatial gap sufficed to indicate the spiritual separation of the two figures. Another frequent method of representing the Annunciation in middle Byzantine churches is especially interesting in this respect. It shows the Angel on the left and the Virgin on the right spandrel of the triumphal arch, the two figures being separated by the wide opening of the arch (2). In this way a large part of the church is included in the image. In the Nativity, the concave landscape with the open cave in the centre is adequately fitted to the physical cavity of the niche in which it is placed (10b, 11). The adoring Angels bow in the most actual sense to the Child who, in the centre of the composition, is sheltered and surrounded by all the other forms and figures. The Presentation in the Temple furnished the opportunity to oppose two groups of figures right and left of the centre, which again is occupied by the Holy Child (14). And in the Baptism the angle of the niche made a fitting frame for the river Jordan (12). Only in this spatial setting does the bell-like or conical shape of the river take on an air of spatial reality: the Angels and the Baptist are really standing on its opposite banks, separated

23


by the gulf of the enclosed space. Moses and Elijah, in the Transfiguration,

could be made to turn towards the Christ in veneration without abandoning their semi-frontal attitudes. The arrangement of the figures has a spatial reality, with the main personages displayed in a semicircle and the three Apostles below

and in front. A diagram of the actual places occupied by the figures in the spatial receptacle would almost coincide with a diagram of the reconstructed

event in physical space. The scope of this feeling for the reality of space can be gauged by the tact

that scenes which called for frontal representation and did not include opposed and correlated groups, or those which, if spatially reconstructed, would take the

form of a row of figures posturing or passing by on a narrow stage, were hardly ever set in deep niches. In these cases such a spatial setting would have been

meaningless; and this is probably the main reason why we never find the Entry into Jerusalem, the Crucifixion or the Anastasis (13) depicted in the deep angular squinches of the naos. These scenes, demanding flat emplacements/1 were

shifted either to the transepts or to the narthex. But there is another, and purely formal, consideration which may have played its part in excluding such themes as the Crucifixion and the Anastasis from the cycle of the deep niches: the long straight lines of the Cross, which appears in both scenes (as also in the Descent from the Cross), would have looked broken or contorted had they been

displayed on a strongly curved surface.42 It is one of the aspects of the Byzantine conception of forms in space that forms were taken at their face value: a line which appeared broken or curved owing to secondary optical effects was taken

by the beholder to be really angular or curved. Lines, on the other hand, which were to be taken as straight, because they were meant to depict straight objects,

could not be placed in situations which would destroy their appearance of straightness. The impression made on the beholder was, in fact, the ruling

consideration in Byzantine monumental art as much in its formal as in its spiritual

aspect. It was-his standpoint and the position from which he would generally

view it that determined the compositional arrangement of the picture. And as

the main viewpoints lay, for liturgical reasons, along the middle (west-east) axis of the building, the compositions were shaped and placed with express

regard to this axis, both thematically and formally. This becomes especially clear if we analyse the compositional arrangement of the festival icons in the

deep angular squinches of Hosios Lukas or Daphni. The main figui e of the sccne Christ or Mary—is invariably represented as turning towards this central

axis, that is, towards the beholder. This is why, in the icon of the Presentation

in the Temple in Hosios Lukas (14),** the normal direction of the narrative, from left to right, was reversed. Had it been otherwise, Mary would have seemed to

be moving away from the beholder, or walking out of the church." As it is,

she comes to stand, in an almost frontal position, with the Child in her arms,

on that part of the angular niche which faces cast. The same is the case in the

24


Baptism in the same church (12a). All the compositions, in fact, are so arranged that the main figures and motifs are placed in those halves of the triangular niches

which stand at right angles to the main axis (the walls facing east and west respectively). The other halves, on the north and south walls, are treated as

secondary'parts of the images, as their "averted sides"-. This is a feature similar to that which has been noticed in the composition of the cupolas: jusi as these had a main view, a "facade", so had the spatial triangles of the squinches. Not only do the main views contain the principal figures; they arc

aLso richer in colour and fuller in form (11). But this was not the only compositional rearrangement which the placing of

the pictures in the deep angular niches of the naos required. A thorough over haul of the models was necessary. Verticals, for instance, which were meant to be seen as such, had to be particularly stressed and specially placed so as not to

appear bent or broken. On the other hand, a seeming curve could be produced

on the curved surface by judiciously placed straight lines. On the whole, the

models were simplified; the painters could do without most of the motifs which, in flat compositions, were necessary in order to establish pictorial coherence:

motifs, for instance, which enveloped and echoed upright figures and groups of figures in order to bind them into a closely woven texture of compositional pat tern. Such patterns, indispensable in the flat, would have had a bewildering effect on the curved surfaces which by their own tension alone performed the function of binding the figures together. The spatial compositions had to be more straightforward than those of contemporary miniature painting. ̂ But the fact that an essential part of the enlivening movement of the composition was achieved, not by the flat design, but by the effects of spatial projection, imparted to these icons a tense and thrilling quality, something of the charm of pictures

in statu nascendi.

In view of the fewness of the niches available in the naos, only 2. certain

number of a more comprehensive cycle of icons could be placed in spatial receptacles. There were no fixed rules to decide which these favoured icons which were to be placed in the squinches and deep niches of the naos should be; but several factors influenced this choice and brought about the final distribu tion. The first was the architecture of the church; for that decided how many

images could be placed in niches. A second factor was the formal character of the compositions; those which lent themselves to being fitted into curved recep tacles were preferred. Theological considerations formed the third and final factor. The rank and dignity of the icons had to be considered, and a certain sequence observed among them in accordance with the liturgical calendar. At length a more or less firm tradition grew out of these determining factors.^ This tradition would favour, for instance, the placing of the Annunciation in the north-eastern angle of the naos, forward left of the beholder, that is, either in the north-eastern squinch of the cupola or in the spandrels of the triumphal

25


irch, beginning on the left with the figure of the Angel. The Nativity would

bllow in the south-eastern squinch. But there the regularity ended: different

mages made up the rest of the naos-cycle in different churches.

The scenes which could not be placed in the niches of the naos had to be

hifted to secondary parts of the interior—the transepts for instance—or to the

iarthex. They were placed, as far as possible, in framed lunettes embedded in

he marble linings of the walls. Crucifixion and Resurrection tended to con-

ront the eastward-looking beholder on walls facing west, in an arrangement

ymmetrical to the central axis. The Koimesis was usually represented above

he main entrance, on the west wall of the naos. In churches which had no

quinches or niches in the naos, the whole evangelical cycle had to be repre-

ented in the transepts; this became the dominating type in the twelfth century

tnd after. Subsidiary cycles were scarce in the monuments of the classical era

»f middle Byzantine decoration: they comprised only the Life of the Virgin

generally in the narthex) and a few Old Testament scenes which were meant

lot as illustrations of a comprehensive narrative but as typological prototypes,

Jid were placed in the side chapels of the sanctuary (42, 43).

§The Choir of Saints

The third and lowest zone of centralized decorations does not contain any

cenic images: single figures alone make up the "Choir of Apostles and Martyrs,

Vophets and Patriarchs who fill the naos with their holy icons" (15).45 These

igures are distributed in accordance with two iconographical principles which

ntersect each other: one that of rank and function, the other that of calendrical

equence. It is the former of these which predominates.^ Sainted priests and

patriarchs arc placed in or near the main apse, in a hierarchical order which

tescends from the Patriarchs of the Old Testament, by way of the Prophets and he Doctors of the first centuries of Christianity, down to the humble priests of

he Eastern Church. The^Martyrs fill the naos, arranged in several groups: the

loly Moneyless Healers (the Anargyroi) next to the sanctuaries, the sacred War-

iors on the pillars and the arches of the central cupola (16a), and the rest mostly

n the transept, distributed in groups according to the dates, of their festivals

n the liturgical calendar. The third category comprises the'holy Monks, who

jre placed in the western part of the church, guarding the entrance of the

larthex and the naos. Holy women and canonized emperors are depicted in

he narthex. But this order is by no means rigid; it allows of variation according

o the dedication of the particular church and to its architectonic type. One

nain variant is to be found in churches whose central cupola rests not on large

quinches (which afforded space for scenic representations) but on smaller pen-

lentives which could be decorated only with single figures. These four single

igures *.re invariably the four Evangelists, seated on thrones in the act of writing

26


their Gospels- The narrow arches which support the cupola in churches of this

type arc filled with medallions of the ancestors of Christ or of holy Martyrs.

But variable as was the arrangement of the single figures in Byzantine

churches, there were nevertheless strict rules of aesthetics governing the manner

in which they were represented in keeping with their different architectural

settings and iconographical significance. The figures can be divided into four

main groups: seated figures^ full-length standing figures, busts, and'medallions. Seated figures occur on curved surfaces only: the only persons represented as

single figures in seated postures arc Christ, Mary and the Evangelists.*6 The

first two (in apses and so on) are always fully frontal, facing the beholder. The

Evangelists, on the other hand, are invariably rendered in an oblique position,

a three-quarter view which is a formal corollary of the oblique position in space

of the spandrels on which exclusively they appear.47 It is in accordance with the

Byzantine feeling for the implications of position in real space that the figures

should express, and to some extent correct, by their projectional attitude their

emplacement in physical space with regard to the beholder. Frontality is con

ceived with regard to the main axis: so that if the surfaces 6n which the figures

arc painted are oblique to this main axis then the figures have to be turned so

as to appear frontal with relation to their real position in space. The projec

tional obliquity of the figures corrects the oblique position of the diagonal

spandrels (16b).

The standing figures find their proper places on vertical walls, in barrel

vaults and on the vertical parts of arches. One main rule applies to these

figures: "standing" is taken literally, that is, they are thought of as standing

vertically in physical space; they are so placed that they stand with their heads

up and their feet pointing downwards. Any other arrangement would be felt

to violate the strict Byzantine feeling for the reality of space. Consequently, we

never find at Byzantium figures which are meant to be standing depicted in the

horizontal axis of a barrel vault or in the apex of some other vaulted roof—a

peculiarity often found, of course, in Romanesque wall paintings or even in

Italo-Byzantine mosaics, such as those of St. Mark's in Venice.*8 Such figures

would have appeared to the Byzantine beholder, who considered the actual

position of the figures in physical space, not as standing up, but as lying. The

vertical position of standing figures, however, was not enough by itself to satisfy

the subtle and strict aesthetic code of middle Byzantine decoration. The surfaces

on which such figures appear must, in addition, have a clear connection with the

floor of the church, the standing-place par excellence. Standing figures, therefore,

are usually excluded from closely framed units which arc clearly cut off from the

floor; for example, receding niches rising from shelving horizontal bases, or

"hanging" spandrels. Such emphatically framed elements would have been

felt to isolate the figures too much from the common standing-ground. Figures

so placed would not give the impression of belonging to the great choir of

27


Saints which fills the church. In a higher degree than of any other parts of the decoration, it can be said of the Saints, the single figures, that they share the space of the church with the beholder. They are closest to him, both actually and spiritually; they arc his ideal representatives in the hierarchy of the Kosmos. Needless to say, the scale of the standing figures and the height of their emplace ment vary in accordance with their rank within this hierarchy.

Full-length figures-stand in many cases on a narrow strip of ground which is sometimes just a band of variegated dark colour, sometimes a meadow with plants and flowers, and sometimes a suppedancum or footstool with a special shell-shaped pattern. These three different designs of the ground-strip correspond to three different degrees of sanctity, the last of the three being reserved for Christ, Mary and Angels. The shell pattern on the suppedancum has a long history. It derives from the shell-shaped pattern of pavements, which for technical reasons, became the rule in the setting of the ground of mosaic floors from the later fourth century onwards. From this it became, sail later an abridged symbol for "floor", standing-ground par excellence, and was accorded to figures of special importance. In the course of the twelfth and thirteenth centuries the original significance of the pattern was gradually for gotten but not its habitual connection with especially venerable figures; it was used, accordingly, as a flat background (instead of a golden ground) for figures

of the Virgin, and so on. Standing figures are mostly represented frontally, unless they are meant to

be members of an interrelated group, as is sometimes the case with Angels who, even as isolated figures, are often thought of as connected with Christ, Mary or other objects of worship: the Angels are, for this reason, often rendered as turning and bowing towards some object. Other exceptions are the rows of Apostles, in apses like those at Cefalu and Monrealc, or in the narthcx of Hosios Lukas (19 17) where the twelve single figures of the Apostles, deployed on dif ferent pillars and arches, are gathered, as it were, into a homogeneous group by the simple means of slight turnings and quarter views. It is important to note

that all these movements point towards the image of the Pantocrator above the main door, as a common centre. It is by these subtle deviations from rigid frontaiity that the Apostles are singled out from the numerous group of other figures (all frontal except the Angels) and related to each other as well as to

the common centre. .

Next in dignity come the half-figures or busts. These appear exclusively on vertical parts of the architecture which are definitely framed and have a strongly emphasized base, by means of which the figures are cut off. Lunettes, shallow and low niches, and parts of spandrels are the ideal places for figural representa

tions of this kind (18a). They are never found on large surfaces which have no definite delimitation at the bottom. Half-figures floating in the golden ground do not appear in middle Byzantine painting., except for those of Angels in scenic

28


representations (the Crucifixion, etc.), where the absence of the main part o the body has a theological significance, as stressing the spiritual nature of thes. heavenly beings. The portraits of the holy Stylites, their upper bodies renderec as emerging from a column or as planted on it, arc half-figures of a specia kind this abbreviated mode of representation was doubtless fostered b> reminiscences of sculptured busts on columnar socles. In the formal arrange ment the figures of the Stylites are treated like full-length figures: they are use. on narrow vertical strips of walls or, for preference, pillars (49).

The remaining parts of the building, especially cross-vaults spandrels

narrow horizontal strips of wall, etc., arc usually filled with medallions (x8b). The triangular segments of cross-vaults, for instance, never show any other kind o; figural decoration. It was felt, apparently, that a bust, whose_spatial realm was softened, as it were, by a circular frame, and which did not imply the con ception of ̂ standing" in real space, was the only kind of figural motif whicl could satisfactorily be placed in the nearly horizontal parts of the vaulting. Tht circular shape of the medallions needed, on the other hand, the support of ai architectonic or linear framework, into which they could be fitted so that thc> did not seem to be rolling or floating insecurely. This is why medallions an hardly ever found on large unbroken walls, and are always enclosed in narrow strips, in triangles, spandrels and such like. In the summits of arches medallion: were arranged so as to have their heads pointing towards the beholder; thus the* appear erect and help to fix the main point or axis of view. The heads generall; point, therefore, towards the centre of the naos, that is, towards the centre o The square which is surmounted by the main cupola. In the summits of later* arches they point either towards the middle axis or westwards, facing th-beholder. The former is the case when the ground-plan of the church is stneth centralized, the latter if it is to some extent longitudinal.

In the strict system of Hosios Lukas the half-figures of the medallions ar strictiv frontal, as with full-length figures, and with the same exception that c, \nzels whose "adoring" attitude was an iconographically fixed formula (19,^7 .

At Daphni, on the other hand, strict frontaiity has ceased to be the exclusive!;, dominating principle. Slightly oblique views of the faces were introduced, a befitted the new conception of larger decorative units to be treated as pic tures" At Daphni three or more medallions make up a vivid group. Possib\) this loosening of the firm structure, which took place at a comparatively earh date was helped by the fact that the medallion as a mode of representation wa. an integral part of the Hellenistic inheritance which had to some extent with stood the influence of the strict mediaeval system and readily reasserted itsel

when this influence showed the first sign of relaxing.


FORMAL UNITY

In the three zones of the Byzantine church building the decoration is dif

ferentiated as regards both the idea of action expressed in the images and the

conception of the images themselves. An eternal and holy presence is manifest

in the paintings of the highest zone, to the suppression of all narrative and

transient elements. There, the timeless dogma is offered to the contemplation

y of the beholder. The decoration of this zone may be said to consist neither of

j" pictures", in the normal sense of the word, nor of single figures. What is

i given is new and specifically Byzantine compositional units in physical space, not

'depicting but constituting magical realities: a sacred world, beyond time and

causality, admitting the beholder not only to the vision but to the magical

.presence of the Holy. In the middle zone the timeless and the historical

elements are combined in accordance with the peculiar character of the festival

icon, which simultaneously depicts an historical event and marks a station in

the ever-revolving cycle of the holy year. In this zone the pictures are so com

posed as to offer a flat image to the beholder while not losing the character of a

spatial receptacle for the holy figures. The pictures are enriched- with single

traits which help to complete the narrative. The Nativity, for instance, often

contains secondary scenes, such as the Journey and the Adoration of the Magi,

the Annunciation to the Shepherds and the bathing of the new-born Child.

But these details are not arranged with regard to continuity in space or time.

Their relation to the main theme is in thought only; they complete the principal

scene inasmuch as they include all the motifs connected with it in the Gospel

narrative.49 Isolated as holy icons and, at the same time, related to then-

neighbours as parts of the evangelical cycle, the paintings in the second zone are

half picture and half spatial reality, half actual scene and half timeless repre

sentation. But in the lowest stratum of the church, in the third zone, are found

neither narrative scenes nor dogmatic representations. The guiding thought in

this part of the decoration—the communion of All Saints in the Church—is

realized only in the sum of all the single figures. They are parts of a vast image

whose frame is provided by the building of the church as a whole.

§ Space

The images of middle Byzantine church decoration are related to each other

and welded into a unified whole not by theological and iconographical concepts

alone, but also by formal means which create an all-embracing and homogeneous

optical unity. The optical principles used for this purpose aim, broadly, at

eliminating the diminution and deformation of perspective. The most obvious

of them is the "staggering" in size of the images and figures according to their

30


height or distance from the beholder's viewpoint. There is some difficulty, ii is true, in appraising the part played by this anti-perspective optic in the general staggering of size, which extends from the lower reaches upwards to the zenith of the cupola, because the staggering was in large degree due to the existence ir Byzantine iconography of a direct ratio between sanctity and height as well as between sanctity and size. Furthermore, several different scales and proportions exist within the variegated system of the decoration: one for festival icons,

another for standing figures, yet another for medallions, and so on. The optical

differentiations due to anti-perspective devices can be clearly ascertained, there fore, only within each of these categories separately (15). The modern beholder, it is true, has the feeling that the ratio of "stagger" throughout the church is some what excessive and that the pictures in the upper zones are certainly larger than

he would expect; but this feeling is too vague to afford a firm basis for conclu sions about the part played by the specifically anti-perspective enlargement.

Fortunately, however, there are cases in Byzantine decorations where images of equal importance and dignity, belonging to a single homogeneous cycle, were placed one above another for reasons of space. Any differences of scale between the parts of such a cycle can have nothing to do with iconographic differentia tion, and must be explained as an optical device. This device was used above all in narrow parts of the building, as, for instance, on the walls of the side chapels of Monreale (Lives of SS. Peter and Paul), which cannot be seen from a normal angle but only from almost immediately below. If the walls were divided into layers of equal height, the upper ones would appear to be smaller narrower than the lower. Staggered as in reality they are, they appear to tht beholder as of equal size (20). This expedient is used in narrow spaces only; bui

when a long-distance view is possible and the effects of foreshortening do not

occur of necessity, then the superimposed layers retain their equal heights. The Byzantine principle of staggering the sizes of superimposed layers for

anti-perspective reasons was not practised in the West. There, walls were cithc: divided into layers of equal depth or consisted of strips arranged in accordance

with an organic idea of growth, with the heaviest and broadest layer at the bottom, the lightest and narrowest at the top. Here we meet with a funda mental difference between the two conceptions of decoration: the Western

architectonic and organic, illustrating the ideas of growth, weight and poise: the Byzantine optical and hierarchical, with the primary aim of preserving the

essential size of the image from optical distortion. But the tendency towards eliminating the effects of perspective foreshortening

goes further than mere staggering of size. The proportions of single figures were also affected by this process (21). Middle Byzantine mosaic figures are often described in studies on Byzantine paintings as "elongated", "ascetic", "emaciated", amd so on. They have, indeed, all these characteristics (from which modern authors too often proceed to draw conclusions as to the "psychology of ?tyle"), if they

31


are viewed or photographed from a level from which they were not meant to

be seen—figures in cupolas from high scaffolds, for instance. In reality, how

ever, the proportions of these figures are in tune with the point of view of the beholder. Seen from below they appear in normal proportions—that is, they

appeared so to the Byzantine beholder who, from what we know of his reactions, must have registered the optical facts in a more straightforward way than the modern spectator, who is apt to see more analytically and to correct perspective distortions automatically if he has a chance to measure distances and angles.50 In Byzantine decorations the painters themselves anticipated the distortions which would appear to the view from below and corrected them by elongating the figures accordingly. They also adjusted the proportions of the figures with regard to the curved surfaces on which they were painted. Thus the legs and lower parts of figures in cupolas or in vaults were elongated more than the upper

parts, because .the upper parts, being placed in the more curved sections of the cupolas or vaults and consequently on surfaces which stand approximately at an

angle of ninety degrees to the optical axis, suffer less from distortion than do the lower parts in the more vertical outer ring of the cupola or the more vertical parts of the vault. In this connection it is interesting to note that the seated

Apostles in the Pentecost cupola of Hosios Lukas show hardly any of these pre ventive distortions. The reason is that they were not meant to be seen from immediately beneath, being placed just above the sanctuary, which was gener

ally inaccessible to beholders; they can be seen from further back (west) only, at a flatter angle, and there they present themselves in a normal undistorted view. That this cupola really belongs, optically speaking, to the naos is also evident from the fact that it is surrounded by heavy framework in the east,

south and north alone, and remains "open" to the view only from the west.

Unbroken golden ground guides the glance of the beholder smoothly from west to east, without any caesura which would indicate a change of the viewpoint.

In apsidal mosaics another kind of preventive distortion is to be found. Figures on the outer edges of the semicylindrical niches of the main apses, for instance, would appear to the beholder who can see them from the centre of the church only (a nearer view being barred by the iconostasis and the inaccessi bility of the sanctuary) much reduced in breadth and therefore abnormally slender. To counteract this effect of foreshortening, such figures^ were made

broader and more "thick-set" than their neighbours which could be seen at a

more normal angle (22) .61 In cases in which it was iconographically possible and which admitted of both distant and near views, the corrective broadening of figures on the edges of semicircular niches was effected by broadly expanding postures, which give the additional breadth required without spoiling the normal appearance of the figures for the near view. Hosios Lukas offers striking examples of this solution, in the two semicircular niches of the narthex, with the Washing of the Feet and the Doubting of Thomas." The figure (St. .Andrew?)

32


at the rieht edge of the former is represented as walking into the picture from the right, in an ample stride, with garments spread out in breadth and accentuated bv broadening motives round shoulders and hips. The Apostle (St. Luke) at the left edge of the niche is depicted in the act of unbinding his sandals, resting one foot on a chair and thereby assuming a very broad pose (23) In the Doubting of Thomas the figure at the extreme left is made to look broader than the pre ceding ones by the simple expedient of showing it whole, whereas the others arc

partially hidden by the backs of their neighbours (24). The broadening of figures on the edges of apses which can only be seen at

an oblique angle occurs once again in the development of monumental painting, namely in the Italian High Renaissance, when similar concepts were revived probably, not without knowledge of Byzantine solutions. One of the chief oi These later examples is the sketches by Sebastiano del Piombo for_ the fresco decoration of the first chapel to the right in San Pietro in Montono in Rome.11 The tradition may in this case have been handed down by Venetian art.

Other and more complicated oblique preventive distortions are to be found in the figures and faces on the barrel vaults of sanctuaries, like those of the Angels in the sanctuary of the Haghia Sophia in Constantinople." These fieures look badly distorted from the near view (25), but quite normal and undis torted if viewed at an oblique angle from below. Even the technical processes took account of the "correct" point of view; the cubes, for instance, which make up the mosaic tympanum above the main door in the narrow narthex ol Haghia Sophia, visible only at a steep angle from below, are set obliquely into the wall so as to have their surfaces standing normally (at an angle ot 90 degrees;

to the optic axis of the spectator." , . .

Underlying these corrective distortions and technical processes is a principle closely akin to the dominating principle of middle Byzantine iconography: the picture is drawn into a close relationship to the beholder. Every care is taken that it should appear to him undistorted; and a very subtle system of perspective lore was evolved and used to achieve this aim. This lore, however, never became a systematic science as it did in the Western Renaissance; it always remained a purely artistic and formal principle without any codified laws, It did not aim at rigid consistency, but it had the effect of establishing formal umtv throughout the main parts of the decoration. Byzantine "perspective might

be described as "negative" perspective. It takes into account the space which surrounds, and is enclosed by, the image and which intervenes between the image and the beholder; and it aims at eliminating the perspective effects ot this space on the beholder's vision. The Western artist, by contrast, subjected his figures to the laws of perspective in order to make them appear as real bodies seen from below, with all the distortions which this view brings about. He created an illusion of space, whereas the Byzantine artist aimed at eliminating the optical accidents of space. The result of the Western practice is a picturo

33


of reality; the aim of the Byzantine artist was to preserve the reality of the

image. The Byzantine image escaped the casualness and haphazardness of

spatial projection; it always remained an "image", a holy icon, without any

admixture of earthly realism. By eliminating the perspective effects of space the

Byzantine artist was able to bring out the emotional values of height and remote

ness in their purest form. The beholder of Byzantine decorations, who sees the

image undistorted in spite of the great height at which it appears, feels lifted up

to its level, high above the ground, to which the beholder of Western decorations

is firmly fixed—pressed down, as it were, by the "frog-perspective" of the di

sotto in su. In this way, not only spiritually but optically also, the Byzantine

beholder was received into the heavenly sphere of the holiest icons; he partici

pated in the sacred events, he was admitted to the holy places of which the

icons are the magical counterparts. Moving along the main axes of the church

and viewing thence the holy icons above him he was able to perform a symbolic

pilgrimage. He was not fixed to one point like the beholder of Western decora

tions, whose illusionistic and perspective constructions seem to "stand up" from

one particular viewpoint only; as he moved along the liturgical axes of the

church the rhythm of the cupolas and vaults moved with him, and the icons

came to life without losing their iconic character.

Only those parts of the building which the Byzantine beholder was not

allowed to enter were presented to him as fixed projections. But even these

fixed projections are connected with the rest and take part in the movement.

The beholder's glance is led from one part to the other, round the conch of the

apse and down the semicircular walls. Everything in the vaulted parts of the

building is gently connected with its neighbouring forms, no hard limits or

accents stop the wandering gaze.58 Even the edges are rounded so as to lead

the eye from one wall to another. That this is no mere impression of the modern

beholder projected into the past, but a conscious achievement of Byzantine

decorative art, is made quite clear by those documents of contemporary aesthetic

appreciation, middle Byzantine Ekphraseis." These Ekphraseis stress again and

again that the glance of the beholder is not to rest on one part of the decoration,

but must wander on in ever-changing directions. The very words of these

descriptions, in verse and prose alike, suggest a dynamic movement round the

centre of the church. Gregory of Nazianzus speaks of the church built by his

father as "a temple returning into itself on eight columns", as "radiating down

wards from the cupola" and as being "surrounded by ambulatories". Proco-

pius, in his description of the Haghia Sophia, underlines the fact that the eye

of the beholder is not allowed to rest on any of the parts of the church but as

soon as it has settled on one is attracted by the next. Photius, in his Encomium

of the Nea, is even more explicit: "The sanctuary seems to revolve round the

beholder; the multiplicity of the view forces him to turn round and round, and

this turning of his is imputed by his imagination to the building itself."

34


There is another point on which the contemporary Ekphraseis offer elucidat

ing comment. The descriptions of pictorial decorations are couched in terms

which suggest the presence in reality of the scenes and persons depicted. The

Encomiasts did not write: "Here you see depicted how Christ was crucified",

and so on; they said: "Here Christ is crucified, here is Golgotha, there Beth

lehem". The spell of magical reality dictated the words. This magical reality

of the decoration which, formally speaking, expressed itself in the spatial character

of the whole and in the life imparted to it by the movement of the beholder in

space, cannot be rendered satisfactorily by photographs. Byzantine church

decorations reveal their supreme qualities only in their own ambiente, in the space

in and for which they were created.

flight

To this spatial ambiente also belongs the actual light. Just as the Byzantine

decorator did not represent space but made use of it by including it in his icons,

just as he took into account the intervening space between the icon and the

beholder, so he never represented or depicted light as coming from a distinct

source, but used real light in the icons and allowed for its effects in the space

between the picture and the beholder's eye in order to counteract its disturbing

influences. The first resource is illustrated by the inclusion of shining and

radiating material in the picture, especially gold, which is so arranged as not

only to produce a rich coloristic effect but also to light the spatial icon. The

deep niches under the cupolas of Hosios Lukas and Daphni are effectively lit by

highlights appearing in the apices (4,1 ob). These radiant highlights separate Mary

from die Angel in the Annunciation and they surround the Divine Child in the

Nativity, the landscape background of which is interspersed with gold in those

places where the golden cubes collect the light in the focal region of the niche.

Generally speaking, colours indicating solid forms were interspersed with golden

cubes only in places where the gold is apt to shine owing to the curving of the

surface. This economy contrasts strongly with the indiscriminate use of golden

highlights in the colonial sphere of Byzantine art. In Byzantium proper the

golden cubes stress only the formal and iconographic foci. The centres of icono-

graphic interest and those of formal composition, which in classical Byzantine

art are identical, are stressed by the strongest light. It surrounds the main

figures as with a halo of sanctity. The reflexions shift with the movement of the

beholder and with the course of the sun, but thanks to the nature of the spatial

receptacles of the icons they always play round the main figures. At night, the

light of the candles and lamps creates fitful reflexions on the golden surface from

which the figures stand out in significant silhouette.

It is not only the derivative light of gold which is included and used in

Byzantine icons, but also the more direct light of windows which open in the

35

BYZANTINE iMOSAIC DECORATION

Tand icons themselves, in the cupolas and apses. The effects of this were loticcd—as the Ekphrascis again prove—and reckoned with.58 The same was rue of the light conditions under which the icon is seen by the beholder. Regard vas had to these conditions not only in the emplacement, but also in the shaping md colouring of the icons. Hcsios Lukas provides an especially interesting cx-

imple. An icon of Christ on one of the secondary vaults (26), whose forms would lave been swallowed up by the splendour of the surrounding golden ground lit Dy a neighbouring window, was executed in a singular kind of modelling which, n spite of the almost blinding radiance around, brings out forcibly the vigorous relief of the face. This special kind of modelling recalls the inverted tonality of photographic negatives: the face itself is comparatively dark, with greenish high lights which contrast strongly with the reddish brown of the main features. Yet the highlights do not appear on the raised motifs of the face-relief, in the places where one would expect to find them in normal modelling; they do not belong to the same conception of lighting as, say, the footlights in a theatre. Their arrangement is altogether unrealistic. They are concentrated in the grooves

and furrows of the facial relief—almost, in fact, in those places where one would expect to find the deepest shadows. But these lights delineate the pattern of the face more clearly and forcibly than a normal modelling could have done. It is as if light were breaking through the features, as if it radiated through the relief at its seemingly lowest points. The arabesque of these highlights is strong enough to bring out the features despite the radiance of the golden ground, whose dis turbing effect is further heightened by the fact that only one half of the vault is brilliantly lit, whereas the other is always in dim shadow. Three more images in the same compartment (Mary and two angels) are similarly treated. That the singular conditions of lighting were indeed the reason for this choice of extra ordinary means becomes quite clear if the medallion of Christ already described is compared with a corresponding image of Christ in the other transept, set

probably by the same artist. In this the light is much quieter and the artist could, in fact did, employ the normal positive modelling with the shading of the

face from light to dark (27, a, b). # Quite apart from its practical value as clarifying the form, the inverted

modelling of the first of these two icons produces a peculiar effect which, at least to the modern beholder, has something mysterious and even "magical" (in the modern sense of the word) about it. Later Byzantine painters were quick to perceive and employ this effect. The inverted modelling with flashing high lights played indeed an important part in Pateologan and especially in Russian painting." As happened so often in the course of the development of Byzantine art, a particular practice, employed by earlier artists only under particular conditions, was later seized upon as a means of conveying a certain expressive quality—a by-product originally, but now sought after for its own sake. A

purely optical technique became a vehicle of expression.

36


§ Colour

In other cases the artists regulated the effect of too* much or too little light by the use of darker or lighter shades of colour.60 Although the basic colours

were more or less fixed by iconographical rules,*1 there was always ample room

for differentiation in the shades actually employed. In choosing these the artists

took the spatial conditions into account in order to achieve clearness of form and thought. The general arrangement of tones follows the partition of the building

into three zones. Thus the lightest hues are to be found in the uppermost zone

of the cupolas and high vaults, where tinged white and gold preponderate. The light colours correspond to the idea of immaterial heavenly splendour, but they

are at the same time necessary in order to bring out the modelled forms at the

great height, surrounded as these are by the colour-destroying, glittering gold.

Darker shades would tend to produce mere silhouettes, opaque islands in the sea

of gold. The next zone, that of the festival cycle, admits of a greater wealth

of colour. Even so, the colours here have a bloom and lightness, and dark and

heavy tones arc excluded. These tones do, however, dominate the lowest layer,

that of the Saints, where dark brown, dark green, deep blue and violet make up

the main scale, fitting in well with the tones of the marble lining. Thus the arrangement of colours helps to underline the hierarchical structure of the whole

decoration.

§ Modelling Technique

The technique of modelling is subject to a differentiation resembling that

observed in the realm of colour. Writers on Byzantine art have repeatedly

pointed out that several different "styles" (that is, different techniques of modelling) are frequently to be found in a single and otherwise homogeneous

mosaic decoration, some of the images showing continuously graded modelling

while others are accentuated by hard and sharp contrasts which break up the

draperies and often even the faces of the figures (28). Images of the latter kind

seem to be composed of splintery, disconnected fragments of forms; the bodies seem deeply furrowed by cuts and dark holes. It has been surmised that this style belongs to an earlier phase in the evolution of Byzantine art, that it is "older"

than the graded modelling and that figures executed in this way are either earlier in fact or at least deliberately "archaizing". In truth, however, the two

"styles", with all the intermediary forms, are strictly contemporary as regards

not only historical date but also stylistic phase.

The application of different techniques in one single decoration was made

necessary by the great differences in size between the single icons and the great

variety of optical conditions. Graded modelling with cubes of given sizes and

of a limited gamut of tones is technically and aesthetically advisable only within

d 37


certain, not too wide, range of size. There are difficulties, for instance, when

tic forms are under "normal" size, because in this case there is no room for

.eploying the whole necessary range of gradation. To solve this difficulty the

aiddle Byzantine artists of the classical period had recourse to a peculiar tech-

ique of alternating light and dark cubes within one single row, achieving by

his chessboard-like mingling of light and dark the rapid transition from light to

aade called for by the smallness of the forms."2 This could be done succcss-

jlly, however, only if the- image so treated was not accessible to close scrutiny

nd if the mosaicist could rely on the effect of the mingling of the alternating

Dnes in the eye of the beholder (29).

Although the mosaicist could thus solve the problem of undersized forms by

hortening the range of tones, he could not use the opposite method in the case

i" oversized images. The limited scale of tones at his disposal did not permit

»f the modelling of large-scale forms—a fold, for instance—by means of a con-

inuous succession of shades. He had only four or five different shades of a

ingle colour. If more than two rows of each shade were required to fill the

Jlotted space, the effect would be one, not of continuous modelling, but of a

'striped" surface (30)." If, on the other hand, the mosaicist used the normal

uccession of one or two rows of each shade, he would quickly have exhausted his

vhole range of shades near the edges of large-scale forms, and in the middle parts

le would be left with a uniformly light (or, as the case might be, dark) tone.

The resulting impression would have been one of a very fiat relief intersected by

harp and narrow grooves. The artist could, it is true, make the single com

ponent forms, such as folds, as small as possible, and this was actually done in

he late twelfth century, as in the large half-figure of Christ in the apse of

vlonreale (31).64 But this invariably entailed a loss of monumental greatness.

The total figure, which is now beyond question more than lifesize, appears to be

rovered with, or composed of, a minute network of folds, or rather wrinkles,

vhich look petty. The organic relation in size between the total and the single

ronstituent forms is lost. So the mosaicists of the classical period found quite

mother solution for the problem. They achieved the effect of plastic relief,

-vhich could not be obtained by normal and continuous modelling, by the

ibrupt juxtaposition of contrasting shades of light and dark without any attempt

it continuity. In this way they effected a treatment of the surface which, seen

rcm a distance, especially in the scanty light of a Byzantine church, suggests a

-elief more powerful than could be produced in the normal way. Ultimately,

tf course, this practice was rooted in late antique illusionism. It demanded

xom the beholder a faculty of seeing in an illusionistic way, of himself supplying

he transitional shades and so connecting the contrasting tones with each other.

Although at the beginning this technique was a more or less optical device,

t took on an iconographical significance in the course of time. The reason for

iiis lay in the fact that the figures so treated (that is, the largest and most distant

38


images in the cupolas and the apses) were at the same time the most sacred, thanks to the parallelism between remoteness (in height or in distance from west to east) and sacredness. But the holiest, largest and remotest of these images were, of course, the Pantocrator in the zenith of the cupola and the Virgin in the conch of the main apse, and so it came about that the abrupt manner of modelling was gradually associated iconographically with the holiest images-more especially with those of Christ and the Virgin. From the twelfth century

onwards these images retained their abrupt modelling even when there was no

technical or optical need for its application, even, that is, in small portable icons and in miniatures. That the abrupt modelling was originally an optical device in spadal decoration was gradually forgotten, and it became more and more a

mark of distinction, of sanctity—a process probably fostered by the abstract and anti-naturalistic character of the broken forms themselves. For these forms were apt to be filled with the expressive meaning of unearthly holiness. If in Byzan tine art forms offered themselves to expressive interpretation, we may be certain that the opportunity was accepted, especially if it favoured those spiritual quali ties towards which all religious art in Byzantium tended. It must be added that, on Byzantine soil, the original optical significance of the broken modelling was sometimes rediscovered, especially in periods when the sources of antique illu sionism began to flow again, as they did, for instance, in the Pakeologan era. The mosaics of the Kahrieh Djami in Constantinople" (32b, 33) or the frescoes in the Cathedral of the Transfiguration at Novgorod (1379)" ^ examples in which both the illusionistic significance and the expressive value of the broken model ling were combined. It is nevertheless difficult to realize that the golden net work which forms the draperies in Italian Dugento panels is, in the last analysis,

a legacy of Hellenistic illusionism (32a). t m m The translation of an optical device into an element of hieratic expressionism

which occurred in the course of the twelfth century is quite in keeping with the general development of twelfth-century art in Byzantium. Not only did Hellen

istic illusionism cease to be understood: the whole optical system of church decoration was gradually losing its consistency at that time, one of the most

crucial periods in the development of Byzantine art.

39

MOTES

1 J. Saucr, Die Symbolik des Kirchengebdudes u

nd seiner Ausstattung in der Auffassung des MitUlalters,

Freiburg i.

B., 1902.

. 1 F

or a summary of this d

octrinal controversy, see K. Schwarzlose, Da Bilderstrett, a

n Kampj

der griechischen Kirche um Hue Eigenart u

nd Freiheit,

Gotha, 1890; L. Brehicr, L

a querelle des images,

Paris, 1904; N. Mclioransky, "Filosofskaya storona ikonoborchestva", Voprosy FilosoJU,

etc., II,

X9°7>' P-'49 ff-I L. Duchcsnc, " L'iconographie byzantine dans u

n document grcc d

u IX' s.",

Roma e

Oriente, vol.

V, 1912-13, pp. 222 ff,

273 ff.,

349 ff.; A

. v.

Harnack, Dogmengeschkhte,

Tubingen, 1922,

p. 275 ff.;

G. A. Ostrogorski,

"La doctrine des saintcs

icones ct

1c dogmc

christologique" (Russian),

Seminarium Kondakovianum,

I, Prague, 1927,

p. 35 ff.;

Idem, "Die

' crkcnntnisthcorctischcn Grundlagcn des byzantinischen Bildcrstreites"

(Russian with Ger

man resume), Ibid.,

II, Prague, 1928,

p. 48 ff.;

Idem, Studien

zur GcschichU

des byzantinischen

Bilderstrcitcs, Brcslau,

1929; Idem,

"Rom und Byzanz in

Kampfc um die

Bilderverehrung"

(Russian with German resume),

San. Kondak.,

VI, Prague, 1933,

p. 73

ff-I E-

J-Martin,

History of the

Iconoclastic Controversy,

London, 1930

; G. Ladncr, "Dcr Bilderstrcit

und die

Kunstlchrcn dcr byzantinischen und abcndlandischcn Theologic", Zeitschriftfur Kwchcngcschkhte,

III. F., I, vol. 50,

1931, p-1

ff; V. Grumcl, "Rccherchcs rcccntcs sur l'iconoclasmc", Echos

d'Orient, XXIX, 1*930, p. 99 f

f.—What follows here is a very simplified s

ummary of the m

ain

i Christoiogical arguments used in the controversy.

■ » T

his thought can be trace

d back to the writ

ings o

f Gcrmanos, at the e

nd of the seventh

century. Sec Ostrogorski, L

a doctrine, etc., loc. cit., p. 36.

4 Both ideas,

that of magical identity a

nd that o

f vencrability,

had become firmly estab

lished in

one branch of popular religious

art is

the fifth

and sixth

centuries, long be

fore the

beginning of the Iconoclastic controversy.

Sec K. Holl, "

Dcr Antcil d

er Styliten a

m Aufkommen

dcr Bilderverehrung", Philothesia, P. K

U-nat zu seinem

70. Geburtstag, Berlin, 1907, p. 54 ff.

The

I popular belief w

as that the spiritual force o

f the v

enerated Stylitcs a

nd their p

ower to aid w

ere

immanent in

their representations.

This seems to

have been the

origin of the

belief in the

miracle-working power of images.

» The problem is s

imilar to t

hat of representing an action on the stage.

But there the solu

tion is rendered easier b

y the fact that the figures arc in m

otion.

• The more recent bibliography on this subject will b

e found in the article "

Kreuzkuppd-

kirchc", by W. Zaloziccky, in

Wasmuth's Lexikon

der Bcatkunst, and in

various papers by N.

Bruncv'(£>:. £«£fcA/i/i, 27, 1927, p. 63 ff; 29, 1929-30, p. 248; 30,

1930, p. 554 ff., etc.).

_ 1

7 Few things, i

ndeed, have kept their f

orm so perfectly a

nd unchangingly as t

he ByzanOnc

cross-in-squarc church. An analogy from a different field m

ay illustrate this stationary perfec

tion and completion: the violin, w

hose shape, o

nce perfected, c

ould not b

e improved upon.

Its form is not affected

by its scale, w

hether sim

ple violin o

r double-bass, just as the f

cnn of the

j Byzantine church remains the

same throughout its

whole range, from tiny

chapel to

vast

cathedral.

• Sec, for

example, the 18th Homily of Gregory ofNazianzus, and Procopius's description

of the Haghia Sophia in C

onstantinople.

» Sec T. Whittemorc, The Mosaics of St.

Sophia at Istanbul,

First prel. report,

Oxford, 1933,

p. 12; I

dem, Secondprel. report,

1936, p.

10. 10 G

. Millet, Le monastere de Daphni, Paris,

1899.

Idem, Rlcherches sur Viconograpkie d

e I E

oan-

gile, Paris, 1916.

» Sec Simeon of Thessalonikc, in M

ignc, Patrologia Graeca,

torn. 155, col.

338 ff.

« FtpucvoC 6f>x\cn\a*6r:<x)

K«v<rravTjvoTrdX«<dj Urropla baXT}aic<mic#|

koi uvonitf|

Stopta, ed. y

Rome, 1556.

The conch of the apse is the cave of Bethlehem a

nd also the Sepulchre of Christ;

^

87


the altar is the table of the Last Supper; the ciborium signifies Golgotha and the Holy Sepulchre,

»°Mignc, Pair. Gr.y torn. 155, col. 3°5 ff-» ^P- 34°. whcrc the Bcma is identified with Heaven, the Synthronos with the Ascension of Christ; the Prothcsis signifies the cave or Beth-

C CI^ The sources of this interpretation arc quoted in G. Millet, Richerches, be. cit., p. 25 ff. 11 For the Western conception, sec the writings of A. Schmarsow, especially his Kompo-

sitionsgesctzc in den Rcichcnauer Wandgcmaldcn", Rep. Jut Kunstwiss., vol. XXXVII, 1904,

p. 261 ff., and Kompositionsgesetze in der Kunst des MilUlaLters, Leipzig, 1915. ^ " This division of the architectural decoration into horizontal zones is in strict accordance

with Byzantine and early Christian, as opposed to antique, cosmography Sec D Ainalov, ElUnistichtskya osnovy uizantiyskago iskusstva, St. Petersburg, 1900; and Rep. fur Kunstwiss.,

XXY,I'A!9H3eis?nbcrg, Grabeskirche und Apostelkirche, II, Leipzig, :go8. The Pantocrator pro gramme of the central cupola was the result of later changes. Sec N. Malicky, •• Remarque* sur fa date des mosaiques de I'cglisc des Stcs Apotres a Constantinople", Byzantwn, III, 1926, p. 123

' W" E WeTaSl O. Demus, Byzantine Mosaics in Greece: Hosios Lucas and Daphni, Cambridge, Mass., 1931, p. no with bibliography; A. C. Orlandos, Monuments byzantins de Chios, 2 vols.,

Athc"(X 'Ichrmt, "Kahrich Djami", I&estiya RussL Arch. Inst. v Konst., VIII-IX, 1902 and

19 "An echo of the tvpc of Chios is to be found in the apse of Sta. Franceses Romana in Rome. Sec R. van Marie, The Development of the Italian Schoob of Panting, I, The Hague,

I923»* o' Dcmus Die Mosaiken von San Marco in Vcnedig, 1:00-/300, Wicn, 1935, figs. 2, 4, 9, 12. » N. A. Bees, "Kunstgcschichtliche Untersuchungen iibcr die Eulaliosfrage und den

Mosaikschmuck dcr Apostclkirchc", Rep. fur Kunstwiss., vol. XL, 1917, p. 59. »T. Schmit, "Mozaiki monastyrya prcpodobnago Lulu", Sbornik Kharkovskago ^L-Fttal.

Obshchestva, XXI, 1913-14, p. 318 ff-; Idem, Die Koimesiskirche von Nikaia: Das Bauwerk und die Mosaiken, Berlin, 1927, p. 52 ff. O. Dcmus, in Diez and Demus op. cit p. 42 ff.

" For example/ in El Nazar, Quclcdjlar, Quaranlcq, Tchareqlc, Elmale, Anahpsis in Gucurcme; see G. dc Jcrphanion, Us eglises rupestrcs de Cappadoct, Paris, 1925 ff passim.

» For example, In Venice and, as a reminiscence in combination with the Pantocrator

scheme, in the apses of Ccfalu and Monrealc. # ... c u » For example, in Nercditzy, where the central cupola contains a combination of the

Ascension and the Glory of the Pantocrator. See N. Sychcv and V. K. Myasoyedov, Freski Spasa Ncrcditry, Leningrad, 1925, Pis. III-XIV.

» T. Schmit, Mozaiki, be. cit., and Nikaia, be. cit., argues that the main reason why the Panto crator scheme was substituted for the Ascension was the diminution in scale of the Byzantine cupolas from the eleventh century onwards. These smaller cupolas had room for eight windows only between which the twelve Apostles could not easily be placed, whereas the sixteen Prophets fitted the given space. In addition, the increased height of the cupolas suggested, according to Schmit, the incision of the Choir of Angels. This argument, however, is not conclusive, because the Pantocrator scheme was already developed in the ninth century. The reasons for the change seem to have been mainly iconographical, the substitution of the dogmatic symbol

for the Gospel narrative. » Dicz and Demus, op. cit., plan after p. 117 and fig. 1. » Daphni, Nicaca, Nereditzy, Monreale, etc. « O. Wulff, Die Koimesiskirche in Nieaea und ihre Mosaiken, Strassburg, 1903, p. 202 tt.

?ToreximpicfinNereditzy (Sychev, op. cit., PL LXVII) Vladimir (I. Ggb^ Freskomalerei der Dimitrikathedralc in Vladimir, Berlin, 1926), Kiev (Monastery of St. Cynl), Pcrcyaslavl and Staraya Ladoga. _ .

» In St. Angelo in Formis, Torccllo and originally also m St. Mark s, /enicc.

88

NOTES

» A reminder that the festival scenes of the Ascension and the Pentecost, later supplanted bv the PanScrator and the Hetoimasia, originally had thar place in cupolas, can be seen in the fact\haTThe^favourite sites for these two scenes from the twelfth century onwards were he vaul'of the transept or the chapcU preceding the two side apses (Cappadooa, M»tra, the

Totura rKjy) ̂ I»>y CTorccllo, Murano). » As originally in the Nea. This meaning is also preponderant in the figure of the Hodc-

ectria above the northern side apse in the Palatina, Palermo. '• Simeon of Thcssalonike; Mignc, Pair. Gr.t torn. 155, col. 347-3J Dicz and Demus, op. cit., plans after p. 117. »• G. Millet, Recherches, op. cit., p. t6 ff., with texts.

. 27 ff.

torted had it been left in the centre, as was normal in fiat « Diez and Dcmus, op. «7., fig. 5, and ground plan

*i In"Sc' .w^ht^fs41 Ltr^ Pau, were al«, repented in .hi, y (^)f « This is true not only of the pendentives of the mam cupoUs but also of the spandrck of

the typieaUy Byzantine vaio with continuous pendenuves (corb«^vaults) as, for example, in the

l89Vwe"apply this analytical kind of optical perception even to flat projections like photo-

^i r: of Torcello, Kiev, Ser«, e,c.

7 Dr.'j: WUde. Thc.ch.pd, b-.-W by Pie,

™TT^^2^J^*t ******. AprWune ̂ p ,* P£ I and IV. » T. Whittemore The Mosaics of St. Sophia, etc., First Report, op. a., ?1 XI Iff. « A Schmarsow has pointed out the analogy between the firmly stepping rhythm ot

A. bchmarsow;nas_po* ^ ̂ ^ ̂ ^ Wcstcrn cpics of chivalryj 35 against the

* Eastern decorations which recalls the structure of the antique

-O. Wulff, "Das Raumerlebnis des Naos im Spiegel der Ekphrasb", Byz. ̂ , 30, 1930.

P' 5 See the description of the cupola of St. Sophia by Procopius (Wulff, be. cit.)} who

89


in Die had been arranged

^eea -.. T Schmit, "Kahrich Djami", ft-** <«• «<- '906; D. Ainalov, VU^

KX .. <fc. Pis- CCVII ff.

II

Kurvacurcn d« ̂ ^^^ZlLituts in Rom, voL XXIII, H

d die

j. Pennethorne in 1837; but doubts as

(j. Durm, Di* Baton* to Crud**>j£. on the interpretation of the ̂ '" optical illusions seems to have been ̂ The theory, on the other toad, by^wh.ch 't Unties are "temperamental expros:ons ot a mental devices of arusts who realized the <> p ^sJ-fights shy of Vitruvius-s statement

S optical consistency of the dev.ces. The i the Greek refinements^^ Ae:r«ul«

Scholars are not agreed

that the curve are intended to correct

£^4 A^ super5eded-that the .rregu-or ^ ^ formalism and tempera-

^ f mcchanical methods" {op. ««•. resl" d nI 5_8), an<J does not explain

aSon inherent in the optical „ ;n mcdi=val architecture see

ofBrookiyn Ins UK

r4^ the artist how to ensure that

»^ 90

NOTES

« tk* hw<: which teach the monumental sculptor ("Definition of pcn>pect:ve: , • • The laws "*£* bcboldcr, , .

■• ■"■ • *•

pp. I ff., 11 ff. jo be. cii., p. 97 ff-11 toe. cit.t p. 171 ff-

" A -&&£ "I -o-ici di S. Apollinare Nuovo e 1'antico anno liturgico Ravennate",

der

P

sonade, Paris, 1846.

und der Agia

St. Sergius at Gaza. See the description by Choricius, ed. Bois-

»• Van Berchcm and Clouzot, op. «*., p. 159

S -*

of Studium and John of Damascus

clasme", L'AcropoU, I ■9s6. p. 5 f- A tin, et Tan des Iconodastes B«U

, Paris, ffitt^ i«

, MiUno ,9.4.

j ̂ lW <&, tAi»«, Paris, 1910. S and the Zisa. L. Anastasi, Verts xcl o. V. Arata, U.,**** -—— »

s goghie de Kiev et Hconographie In,

Early Muslim Anhituturt, I, Oxford, 1932, p. 229.

91

jl . / V V i. Hosios Luoa-. ("iiilifilicon : Interior.

.1/1:' lhr;-l)cmit.\.

. Torccllo, Calhcdral : Main apse jo/o .Hinuii.

u

i

j. Momcalc, C'a t hrclral : West wall of Snuilinn Transept.

I'lialii Almtiii.

Catliolicnn : Anmimiation :\\v\ Samls.

I'linln . \inum.

,•; Wm.r. Si. M;irk"s : Ascension -upol

a. Homos Lucas. Caiholicon : Virgin in the main apse.

Alter Dici-Pi'min.

ioh. Daphni, Catholieon : Nativiiy.

Pholn Atiitni.

IISRlsM*

•: ■ ■'■■(*■ :'"5iivfe)'&^

<5^*r3»'i; v- m*

i;.\. I losios I,ucas, Catholicon : Anasiasis.

After I)ic--I)emu.\.

\\\\. t'.hios. Nca Moni : AnasiasU.

After Diez-Dnnus.

14. IIf)sios Lucas, C'2iiholi<;<ji) : Prcscmaiifjii.

Daphni, Catholicon : Vaults with figures of SaiiMs.

After Dicz-iJemu.\.

MWi

mm

* .'.'i. -5;

B«& ♦ :KfciWw2w« '^^^ T, *"•■ K AV I /ft

,,, Mo,uvalr. Caihcdr;,! : Scenes IVc.m .l.c Lives ol SS. IV.cr a.nl Paul. /V/(i/<; Aiiilcmiu.

1

1

1

V

2i. Salonika. ll;ig!ii;i Sophia : Apostles from main cupola.

.[i'Ci Dirltl. .'(' I'..•iKiaiu. Siihuli>>.

22.

Torcello, Cathedral

: Apostles in

the main apse.

Photo Alinan.

Torccllo. Cat'r..

.tl

: Apostles in the main apse.

. ■■

'.''

, Alinari.

■2-).

Hosi

os Lu

cas,

Caiholicon

: Washing of th

e Fe

et.

. I fin Dir.z-Dcmus.

\-lhl •{^n fit lunofj W\-

.ip ■iimiA .vk)).v|s.).kI ui piuiy : \;

'•v/i' '0|<l<;i.UHUMSllu; ) >(Ji

ji).

liosios l.uras, Cathohcoii

: Medallion with Chris:.

M'trr /)/>•-/)/■///>/».

...7a.

Hosios Lu

cas.

Calholicon

: Vault wi

th me

dall

ions

. ■2

-jn.

Hosios Lucas, Catholicon

: Vault with medallions.

Afu-r Di,-.z-/)niin>.

'_>?>. Daphni. '"ailiolicon : S. Michael in livnia.

I/;., t)i,::~/hiiui<.

■$2,\. New York, Co

ll. Duvcen

: Vi

rgin

(panel on wood).

A per Murntov.

•i2B.

Cons

tant

inop

le,

Kahrich Djami

: S.

Pe

ter

from Death of the Virgin, detail.

I'lw

lo

'.'chfil

* nnijonillei.

acl on wood

1 • vr Mttrnloi.

t j

from Death of the X'irgin, derail.

Photo Scbnh and Joailler.

:-;;j.

Constantinople, Ka

liri

eli Djami

: Nativity.

I'lw

lo Scbah and Joa

illc

r.

I. IS I 1)1 SA I VI S

1 k: xc:

2 Maty

3 John ilu- Korciunncr

4 (jrtg>.rio> Thcologo-

5 Alhanasin.

6 Thcude.iu< T(on

7 Silvester

S Cypiunos

9 Spymlnn

10 Aeheilios

11 Anlhemios

12 Kleulhenos

13 i'ol)katpcs

14 Anlipas

15 Gregorios ol Nyju

16 Philotheos

17 Hietotheos

18 Oionyiios Atcopagites

19 Ignatios Theoph,

20 Gtegotios of Atmenia

21 Kitylloi ol Alexandria

22 mes

23 IC XC

24 Aiuentios

25 Bikentios

26 ViVloc

27

28

29

30

31 32 Theodoros Titon

33 Nestor

J4 Drtnettios

35 Piokopios

36

37 Mcikuciot

38 Gtotgios

39 Nikohoi N«oi

40 Ihtodotoi Sltililjtcs

41 Maty

42 (Pinlcleimon)'

43 IC XC

44 Uiitl

45 Zichitiis

46 Rjf.ul

47 Jiion

48 Soiipattos

(9 KUopas

50 Ananias

M Maty

52 l.ukas Sni.olc

53 IC XC

54 Michael

55 JaWobos AJtlphoiheoi

56 Gibiirl

57 N>kanor

57a Timolhtci

53 Silas

59 Pcochotoj

60 Stephanos

61 Baimbas

62 jn'in Ih' Fotetunnt

63 NiVolaos

(.4 Gtcgotics ihiuma-

tutgos

65 ttasihos

fc6 Aduanos

67 . -

frS

69

70 Theodosios

71 liulhymlis

JJ Pathomioi

77 |,,

p

VJ Antnnim

S4 Kphmm

45 Atscn.ok

!■(• H,lJn. ,

fi7 N't.Ion

SX Ootolhc,^

hi I hcokliilLi

90 Maximios

91 Snocs

92 Joann.V.os

ly) Theodotos Sluditei

94 Daniel

95 l.ulas Gutnikiolei

96 Agjlhangelos

97 Neophilos

9S NiVel..s

91) Akakios

100 Ititas

101 Hjsiliskns

102

103 i:mon

1CH

105

106 Maty

107 Gafcncl

10S |ohn Ihc Koietunni

109 Michael

110 Panlcleimon

111 Thalila.os

112 Ttyphon

113 MoVius

114 Ko>niJi

115 Kyio.

116 Oamiano!

117 joannci

118 Thomas

119 Kaitholomaios

120 t'hilippos

121 Paulos

122 Jakobos

123 Joannes Theolugos

124 Pelios

125 Matkoi

126 Andteas

127 Matihaist

129 I ukas

130 Thtkla

131 Anaslisia

132 Konstanlinoi

133 Htl.r.c

114 Peltonia

135 Agatha

1J6 Kugenia

137 I'igasios

138 Anempoditos

139 AkinJ>nos

\\C Aphthonioi

HI Klpidiphoni

U2 Irene

113 Euphemu

I. KathM.na

145 Manna

146 Batbata

14/ Juliana

llosins Lucas, Catliolif fin : Ground pi.in.

Afltt Diez-Dtnv--..

" 2

§ "

" 2

3

'■i'tllr\

: l.:

|sl

:fil

-s'i

i.|s

f||r

=.iT

frif

i-ii

l'I:

= 'I

^

::

: 2

""?i?

2 2

S|'S

" '

3 2'

§ x?2 fi ^

3 "if

"" "*

"

LIST OK SAIN TV

) Throdorw Slrjitil.t*

>. lt<b(ho,

1

,M..,l,i,..,

■« l'jni.i<,mo«

10 Sophjnoi

n Juchim

;• Dj«.<!

H luuh

:• Thfodoroi SiuJtlr.

16 TKroJniiDt

51 Simton

32 ...

Styli"*

Daphtii, C'rilholicon

: Clrouiul plan.

43B.

Cliios, Nca Moni

: Ground plrin.

After Die--Drums.

Moinci'.lc, Cathedral : Presbytery

I'luili) Anderson.

byzantine mosaic decoration demus

Documents